AI and Elections: A Conversation with Secretary Steve Simon of Minnesota - Harvard Data Science Review Podcast

Loading summary

Transcript36 lines

[00:02]
Liberty Vittert
Welcome to the Harvard Data Science Review Podcast. I'm Liberty Vittert, the feature editor of the Harvard Data Science Review, and joining me is my co host and Editor in Chief, Shao meng. After the 2020 United States presidential election, there were widespread claims that the election was unfair, insecure, and in some instances, outright stolen. This outcry put election officials under intense scrutiny, particularly those in the Secretary of State offices across the country. As we approach another presidential election season, it's crucial to address the most significant election safety concerns for the 2024 presidential, Senate and House races. Today we delve into these issues with Minnesota Secretary of State Steve Simon. How does election reporting function? How will AI influence this election? Is your vote really being counted? This is a special edition of the podcast as Secretary Simon joined us for a Conversation with Leaders piece earlier this month. Please join us for these discussions and more on the Harvard Data Science Review podcast.
[01:11]
Shao Meng
Well, thank you so much again, Secretary Simon, for accepting my invitation to talk to hdsr. Conversation with Leaders I really enjoyed your presentation in May at the University of Minnesota's Data Science Initial Conference as a statistician and a citizen, because I found your presentation both insightful and informative, just like how you started in your presentation there. Could you share with HDSR's readers and listeners what your office does or does not do regarding the election? And are these things typical for other states or there are considerable variations among them?
[01:45]
Steve Simon
Well, first of all, I want to thank you for having me. Thanks for initiating this discussion. I really appreciate your interest in the topic and I really appreciate the work of hdsr, what I've come to know and understand of it, and congratulations on your recent milestone and I wish you the best. So thank you for having me.
[02:04]
Shao Meng
Thank you, thank you.
[02:05]
Steve Simon
I mean, as Secretary of State in Minnesota, as is the case in almost every state, we oversee the Minnesota election system. And it's that way, as I say, in most states. So here's what that means. I'll start with what it doesn't mean. We do not count votes. And that is the case in almost every other state. The Secretary of State does not actually count or tabulate votes. That happens at the local level in towns, in cities, in counties. It's very dispersed, it's very decentralized. That is work that is done at the local level by local people, often the friends and neighbors of the voters themselves. We compile and aggregate others counts, but we don't do the counting. Similarly, we don't hire or train or pay the actual election workers. In Minnesota, we call those election judges, but they can be called poll workers or election workers. That too is a local function and responsibility. But here's what we do in our office. We knit together those various elements of the system. We provide legal guidance and policy guidance. We certify all the elections equipment, the tabulating equipment for use in Minnesota. We design and run the giant databases that are required by federal law that are sort of the spine of the system. And maybe most importantly, we do the policy work at the Minnesota Legislature and with our federal partners in Washington, D.C. so as a result of all of those responsibilities, I like to say that I and we are in the democracy business, and it is quite a time to be in the democracy business. But I think understanding that setup and that division of labor and division of responsibility is really important for understanding the system itself.
[03:51]
Liberty Vittert
Secretary Simon, I can't help but feel like this is a significant election year and especially with sort of major players and what feels like a new external factor of AI in the mix. Do you see any major shifts from your perspective as Secretary of State? Is there anything you're really looking at?
[04:14]
Steve Simon
Yes, and I think it was embedded in your question. There's no question in my mind that AI related considerations are pretty new in 2020. We were talking about a lot of things. We were talking about the challenges of COVID and a once in a century pandemic. We were talking about election disinformation, as we almost always do. But this is really the first time in a sustained way that we who do this work, we who are in the democracy business, not just secretaries of state, but others in every state who do and oversee and administer elections. This is the first time we've really been focused on the AI issue. And as I like to say, I don't view AI as a particularly new threat in and of itself. I do view it as a new way to amplify and magnify existing threats. The existing threat is, for example, election disinformation. That problem isn't new, it's very old, actually. But this is a means of amplifying and magnifying that problem, that threat, which we haven't discussed now at hdsr, you may have been discussing it for a lot of years, but to lay people to people not immersed in this area, this is the first election, certainly the first presidential election, when those considerations have been on the table. So that feels new, it feels different, it feels bigger, it feels like a new kind of challenge. So I would really point my finger there as the thing that most distinguishes this election cycle from the Ones that I participated in in the recent past.
[05:52]
Shao Meng
So in what ways are you seeing AI's impact on the election? On the ground level, what should the average voters be concerned with?
[06:00]
Steve Simon
Well, it's a matter of how we all take in and react to information of all kinds. And in this respect, as I'm sure you know, given your expertise, the elections area is not so different from any other area. AI and deepfakes in particular can impact the way we perceive information in a hundred different areas of our lives. Election is just one, and it's a timely one, and it's the one that I do for a living. So I don't pretend that this is necessarily unique, but on the ground, it means a couple of things to me anyway. One is it emphasizes the need for close collaboration. We're one office among many that touches this area of our life, this area of democracy. And it's essential that we reach out to partners on this, because so many people have different areas of expertise. I'll give you an example. When it comes to AI, we in the Office of Secretary of State, we aren't the technical experts, we aren't the data experts really, in the sense that when it comes to detection, is this thing, is this phone call, is this video, is this audio clip, is it or is it not a fake in the sense that it was done by generative AI, and it isn't what it purports to depict? Right? We aren't the detection experts. We really leave that to others. But that means we have to partner with others. We have to go to those who can tell the difference. Now, once it's detected, we do have certain areas of responsibilities. It's we who can speak to voters. It's we who can use the channels we already occupy to make sure voters are getting good and accurate information. But to me, on the ground, that's one of the things it looks like. Now, that's not necessarily, as you ask, what voters would see. Right? That's what we who administer elections would see. What voters will see from our office, and not just our office, from many around the country and around the state, is more and more attention paid to media literacy, urging voters to show a healthy skepticism about the messages they receive about the mechanics of elections. I'm not talking about who they should vote for or which political party or candidate. We're not in that business. But in terms of the rules of elections, the who, what, why, when and where, we are going to continue to urge people, especially in light of AI, to really seek out trusted sources on that issue. In other words, don't just trust what's in your social media feed, at least not alone and by itself when it comes to those rules, go to a trusted source. We hope that includes our office. For those who live in Minnesota, they'll go to our website. But if it doesn't, for whatever reason, go to one trusted source. That is something that has integrity to you. It could be a city, it could be a county, it could be someone else. But really, practitioners, people who are administering elections, before you believe something in your Facebook feed or that you see on X or elsewhere in the social media world, go to those trusted sources. So on the ground, what I think in this pre election period, as we're approaching November, it's more and more emphasis on that voters will see invitations for them to be more discerning and discriminating. We want people to have a healthy skepticism, but not an unhealthy skepticism. And where that line is is a difficult one. But we want people to view assertions about the election system carefully.
[09:33]
Shao Meng
Let me follow up quickly how much the education system can help in this process, like universities, even high schools.
[09:42]
Steve Simon
I think it can help a great deal. And in Minnesota, we recently had a debate about that. There had been a long running discussion, finally resolved last year in 2023, about establishing some sort of civics requirement in high school in Minnesota. There was none before and there is one now. Now, civics can mean a lot of things, but what it means at a fundamental level is just very basic things about teaching young people the architecture of government. Who does what, what does a court do, what does the executive branch do? What does a legislative branch do? And maybe on a deeper level, what do people like a Secretary of State do? I mentioned at the outset, in response to your question, one of the things we don't do is count votes, right? People would be less susceptible to some forms of disinformation if they understood that very basic piece of the architecture. And what I'm referring to is the fact that if any of your listeners would do a quick 10 or 30 second Google search, they would quickly see that there are people out there asserting falsely and wrongly that I, and not just me, but my colleagues around the country who have similar portfolios and roles are secretly scheming in this pre election period in some back room with our secret algorithms. And we are plotting and scheming, so they say, falsely, to spike the vote totals for the candidates we like and to depress the vote totals for the ones we don't like. Now that Just isn't a matter of opinion. Someone who understands the basic architecture would then understand, hey, Secretaries of State do a lot of things, and I like some of those things and don't like some of those things, but I know that they don't count votes. That's one thing they don't do. So this thing must be false. That's one very basic example of the way in which some education, particularly of young people, can be helpful and can pay dividends in the long run. People understand who does what in our country, just mechanically speaking, they are less susceptible to these false appeals on social media or anywhere else.
[11:48]
Liberty Vittert
I think that that makes me bring up the question of whether AI, you know, not necessarily disinformation, but could it be affecting the election process at a systems level? You know, that was a huge part of the discussion in 2020, the Dominion Systems, you know, all this kind of stuff. Could AI influence the way elections are run? Or really the security, I suppose, of election data? Or is that also not true?
[12:14]
Steve Simon
Well, as we're coming to discover, at least those of us who are relatively new to this area, you know, AI holds a lot of promise and possibility in many ways. It can have a positive impact and influence. It can make life and processes across the board more efficient and more effective and more surgical and. And all the rest. To that extent, it can be a real benefit. But from a systems perspective, could it introduce instability or mischief or something like that? I suppose it could, which leads to a renewed emphasis in our case anyway, on cybersecurity, for example, a really important topic just to make sure our systems are secure, whether it's the databases we run or the transmissions between us and counties and cities, to make sure there's no level of mischief there. So the good news is I haven't seen an example yet of that kind of mischief or instability, but as I'm sure you and your listeners know, it's always a question of hoping for the best, but planning for the worst. And that really is the name of the game this election season. So whether it's cybersecurity or other means, we're always looking to make sure that from a systems standpoint, things are secure. And sometimes that just means using your imagination to try to predict or figure out what could go wrong. And we're doing some of that with a lot of partners at the federal level and at the state level.
[13:50]
Shao Meng
Thank you, Secretary Simon. Now, minasuda is often seen as a golden standard for election data reporting and accessibility. How does the office ensure the elections are safe, accurate and Efficient, particularly in terms of reporting and other large scale election tasks.
[14:06]
Steve Simon
Right. Well, I mean, part of it is to have a redundancy in place. Part of it is to make sure these things are tested, tested often, tested frequently, tested regularly. And then there's another public accountability piece as well. And I think sometimes this gets overlooked in these discussions, which is to say it's not just about the systems being effective and safe, it's about the perception that they are as well. And what that means to me anyway, and to many of my colleagues around the country, I think they feel the same way, regardless of their political orientation or anything else, is transparency is making sure the public understands what some of these processes are in terms of security. Let me just give you an adjacent sort of example from Minnesota. And I think other states, not everyone, but several, have this same thing. So take the issue of elections equipment, the election tabulators. So there are some that assert wrongly that these tabulators have been, whether in Minnesota or elsewhere, switching votes from candidate A to candidate B. Now, speaking only for Minnesota, one of the reasons that I know that's not true is all the work we do leading up to the election and after the election in looking at the performance of this equipment, one of the things we do before the election in terms of security is we have something in state law called a public accuracy test. Now I know before everyone's eyes or ears glaze over, that sounds like a perfectly bureaucratic exercise, a public accuracy test. Or Pat, I get how that sounds, but here's what it really means. It means under Minnesota law that every jurisdiction, whether it's a county or a city or a town that owns and uses tabulating or elections equipment, they must, within a couple of weeks of every election, not just a general election, a primary, a special election, every election, they have to trot out that elections equipment in a pre noticed public meeting called a public accuracy test. Anyone off the street can come. There are no engraved invitations. You're not limited to the one that's in your area or where you live. You could go to one three hours away if you wanted to. And it's a pre notice public meeting where anyone off the street can come in and watch the elections. People basically try to trick the machines. That's what these tests are. They go in there, they'll put stray marks, they'll bend the paper, they'll circle or underline instead of filling in the oval. They'll overvote, they'll undervote. The point is to try to trick the equipment and only if that equipment performs perfectly, will it be allowed for use in the subsequent election? So that's part of how we sort of stress test these systems and in this case particular equipment. And I think that's been successful. The more people can learn and know that this even exists, they don't have to attend. People have lives and families and jobs and all the rest. They're not necessarily going to go to these. But knowing that they exist, I think can give people peace of mind. And beyond just peace of mind, it actually does improve the performance of the equipment. It's public, but even if it were private, it's a way that you're testing the equipment beforehand to make sure it works.
[17:19]
Shao Meng
Quick follow up question that you said. You know, one thing you need to building duplication duties test. I assume you need to have a lot of people to work on those things. Now we have a lot of listeners, data scientists or AI specialists or whatever they are. And just for their benefit, suppose some of them wants to work for your office or generally for government agencies. What kind of skill sets do they need? In addition to the technical side, what are the other things they should learn to equip themselves in order to contribute to this great process?
[17:49]
Steve Simon
Yes, that's a great question. I don't think I've ever had that question before. That's a really good one. I would say a number of things. One is a belief in the work. As corny as that sounds, I think that that really does count for something. The belief in being in the democracy business. That is important work, it's worthy work. Another one though is in terms of skills. They're the technical data skills, of course, which always apply and are useful and important, but also collaboration and cooperation, not just internally in our office, where we're always trying to build a better mousetrap, so to speak. We're always trying at least to be in a process of continuous self improvement. But also cooperation outside the walls of the office, with other entities, with other partners, both in the nonprofit sector and in the public sector, other layers and levels of government as well. Those are some of the things that we look for. Again, as I said, in this business, as in so many others, we're not unique. It's a question of hoping for the best, planning for the worst, and being creative about what that really means. So it's that creativity, that sense of collaboration. Those are the sort of, oh, you might say softer, intangible sort of skills, in addition to degrees and particular technical expertise.
[19:08]
Liberty Vittert
Secretary Simon, I've heard, you know, different states have different systems and that you're speaking obviously from Minnesota, but do you think that there's a chance the existing confusion from these varying state election systems that AI could make this issue worse for voters? Is there a reason that states should start coordinating to sort of create more uniform election rules and regulations, or is there a really good reason for having different systems in different states?
[19:37]
Steve Simon
That is such a well timed question because it's not only come up recently, it comes up constantly in the sense that we're always trying to figure out what the best division of labor and division of power is. Now, I will say that the rough consensus in this country, with some exceptions, is that the time, place and manner of elections, the default position is up to the states, with huge exceptions. I mean, I'm sure most of your listeners are familiar with the epic and important 1965 Voting Rights act, which really stepped into this space where in light of what was happening in certain parts of our country, Congress stepped in. It upended that longstanding consensus, at least on a somewhat limited basis, and said, yes, our default position is that the 50 states have autonomy in election administration. But in certain extreme cases like this, we're going to step in and we're going to impose a federal rule. So from now on, no, you are no longer allowed to have poll taxes. No, you are no longer allowed to have literacy tests for voting, or so called character tests for voting. We know how that went down in the 60s in certain states in certain parts of our country. So there Congress stepped in and said, I know that's our default rule, but we're imposing ourselves here. Even that is more the exception than the rule. We remain a country of basically 50 plus individual laboratories with some federal oversight and some federal rules, like the Voting Rights Act. You used an interesting word, and maybe I'm reading too far into it, but I give you credit for using it, which is the word coordination, meaning can states coordinate without necessarily Congress mandating something? Because I think that would run into a lot of state opposition, because states tend to want to have that autonomy and that independence and that authority. But coordination implies to me anyway something different, something a little more voluntary, something like finding a consensus that all states can adhere to or move towards. And I think we've seen that in some areas, it may not be in every area, it may not be in an area like voter registration. But when it comes to some of the technical matters that we've been discussing today, I don't see why we can't and shouldn't move towards closer coordination as you put it, which isn't necessarily Congress stepping in and mandating things, although I support some of that too. A good measure of that. But even short of that, states talking more, collaborating more, trying to harmonize their approaches more, particularly in areas of data. Not so much rules around, like I said, voter registration, but data, data used, data disclosure, security protocols and practices. Those are things that don't necessarily require an act of Congress and a one size fits all. But more and more states can learn from one another and maybe over time it won't happen immediately move towards some rough standardization on certain issues like these issues.
[22:48]
Shao Meng
Thank you. Secretary Simon, Besides coordinating or collaborating with other states, what other organization does the Secretary of State's office work with to ensure our elections are secure, fair and well reported?
[23:01]
Steve Simon
Well, the first and foremost organization that I have to mention, just because of the place of respect that it occupies, is the national association of Secretaries of State, or nass. And if I can say, it's the oldest such organization in America, the oldest organization, bipartisan organization of statewide elected officials, founded in 1904. And it is what it sounds like, it's the national association of all the Secretaries. We get together physically twice every year, once in the winter in Washington, D.C. and once in the summer on a rotating basis in various locations across the country. And we talk and we compare notes and we exchange best practices. And even when we're not meeting physically, we have one off or occasional zoom meetings on particular topics where we, and not just we, but members of our staff will collaborate and trade and try to learn from one another. So I can't say enough in praise of NASS as just a great connector between and among Secretary of State offices. But there are other organizations as well that we deal with. We have a great partnership, I think, on security and data issues with the federal government, with the Department of Homeland Security in particular, and even on a more granular level with cisa, which is of course part of the Department of Homeland Security. Just for the benefit of your listeners, we in the elections area took a big step forward in 2017. 2017 was the year when the Department of Homeland Security designated our election systems as critical infrastructure. And some of your listeners are maybe aware of what that means. Before then, I hadn't been familiar with that term, but critical infrastructure was a post 911 sort of department of Homeland Security designation that applies to various segments or sectors of national life. Hospitals, military bases, the power grid, the banks. You get the idea. Key core sectors of our national life that deserve special status and special government help and special Protections and defenses. And in 2017, this became, I believe, the 18th sector of our national life to be given that designation. What that has meant is a locked in, really effective, I think, partnership with a lot of federal entities, some of which I just mentioned. So they've been a great partner as well when it comes to security and data and technology and frankly, intelligence. One of the advantages of having this critical infrastructure designation is that I and a small number of folks in my office, and this is the same in other Secretary of State offices, we get intelligence briefing and security clearances that enable us to do that. So they've also been a great partner as well on all these issues.
[25:52]
Shao Meng
Thank you. It's great that you work with all these different agencies. I want to also ask you about any new initiatives that your office is doing, maybe with some of these partners in the AI space.
[26:06]
Steve Simon
Yeah, well, in terms of AI initiatives, I mean, one of the things, I don't know if this qualifies as necessarily an initiative, but one thing that we started out doing at the beginning of this year back in January, is we convened a group to do a tabletop exercise at a military facility in Minnesota. And we brought in the Department of Homeless Security in cisa, FBI, the US Postal Inspection Service, many state Minnesota law enforcement organizations, and critically, a majority of our Minnesota counties were represented as well. And what we did during this tabletop exercise is we gamed out a lot of scenarios that could affect election administration, including, including very much AI infused scenarios. Now, I'll spare you all the details, I can't give a roadmap to the bad guys, obviously, but I will tell you that we had a lot of intense, lengthy discussions over exactly that. Who does what, what is the division of labor. And so that's a new initiative. It's not so much technical in nature, but it is a level and layer of collaboration that is relatively new to do it, to be focused on this particular issue so that we can all understand who's doing what. That's one initiative that we're undertaking, you know, another initiative, again, it's not so technical in nature is what I discussed a little bit earlier, which is really intensifying our efforts to engage in media literacy to make sure that people are better positioned to take in real information and to reject false information about how our election system works, what it is, and we think that's very important as well.
[27:52]
Liberty Vittert
Is there a way that a citizen, if they are concerned about election security, can see for themselves? Are there parts of the election process that they can go watch or what would be the sort of the options there?
[28:05]
Steve Simon
Yeah, I'm really glad you asked that because I happen to believe that one great way to push back against disinformation or really stop it before it starts is to be as transparent as possible. So earlier I mentioned the public accuracy test as one way before the election in one specific area, that is the integrity of the elections equipment or the tabulating machines. That's one way that we invite scrutiny, we invite public view and participation, and we invite people to ask tough questions and make up their own minds. But maybe I should talk a little bit about what happens after the election. Because it's after the election, not before, of course, that you get allegations of the kind that we've seen in recent years, false allegations, saying, for example, that the equipment has transferred votes from candidate A to candidate B. What we do in Minnesota and that many other states do after the election is we have post election audits or reviews where those kinds of issues are addressed, where we do a formal audit and review of the system, where we reconcile the reported results with a randomly selected precincts all over the state with the paper ballots to make sure that the machines were accurately tabulating and reporting what the returns were. And even then it goes through two layers of review. We have county level review and then statewide level review. And after that, if a particular citizen or person is dissatisfied or suspicious or skeptical or even hostile, then they can go to court and marshal the evidence and try to persuade a judge of their point of view. And that judge has the authority to reopen part or all of a particular election contest. So there are multiple areas or avenues of accountability and review and checking the work. That's what happens after the election. It's not just about before, it's after as well. So we hope overall that that persuades people who are suspicious or skeptical or even hostile that there are multiple eyes on multiple people at multiple levels and that ought to give them confidence in the overall system. That's not to say mistakes aren't uncovered occasionally. But overall, the integrity of the system is good in part because there are all these avenues of accountability. So that's something that I and many of my colleagues across the country have been emphasizing in their states as well.
[30:22]
Shao Meng
It's great to hear that. Obviously, transparency is incredibly important. And Secretary Simon, that we have been talking so far is on the AI in the elections that I hope you don't mind that I'm going to ask you a little bit more personal question. Part of the reason that we run this conversation with leaders Is obviously want to inform the community, the data science community, how leaders are using data, so on and so forth. But we also like to inspire some people like to be a leader some days. So obviously, as a Secretary of State, you do a lot more than just election. So first, if you can tell our audience thinking generally, what do you do in terms of your overall charge and also what are the excitement and the challenges? Just give us the general sense. What is it like to be a Secretary of state?
[31:11]
Steve Simon
Yeah. Well, I'll tell you this. Here's what's interesting about it. So I previously served in the Minnesota Legislature. I was in the Minnesota House of representatives for 10 years. And sometimes I use a food analogy to distinguish those two jobs. Both are leadership roles, but they're a little different. Being in a legislative body anywhere in any state or at the national level is like visiting a buffet. You can load your plate up with dozens of different combinations of things. You can take five big heaps or 20 little spoonfuls or whatever, and there's so many issues you can be involved with. But in the Office of Secretary of State, the portfolio is more limited. I don't do tax policy anymore. I don't do healthcare policy or transportation policy anymore. So it's not a buffet. Rather, it's a fine meal with limited courses, four or five or more courses, delicious. You can dive into the flavors, but it's not a buffet. So what are those courses we mentioned? Elections. That's the one that gets the most attention. But we in our office and almost every other Secretary of state office around the country also has a business services department. What that means is it's the place where businesses start, and it's the place where people do other kinds of filings, all sorts of corporate and commercial related filings. So very often the very first thing that someone will do when they start a business, before they hire anyone, before they get an office, before they get insurance, the first thing they have to do is make the thing come alive legally by making it an official thing, an official entity. And you do that and much more through our office. So we like to say we're the welcome mat for Minnesota businesses. And that's really gratifying work. That's one thing we do. Another one that I'll just highlight in the top three. So elections, business services. And another thing we do is we run a program called the Safe at Home program. Safe at Home is an address confidentiality program for victims of domestic violence or stalking or sexual assault. And participants in the Safe at Home program get to have their home address, their true physical home address, kept completely secret and private for all purposes, whether it's interacting with a public entity or a private entity, like a bank or a retailer or anyone else. So the only one who knows where the participant lives is the participant, our office, and whoever the participant chooses to tell. And so we perform a lot of services for them, from getting them to vote safely to handling their mail in a safe way that's not going to reveal where they actually physically live and so on. And it's gratifying and really life saving work that I'm. I'm so really privileged to be a part of. And it has saved lives, there's no question about that. So those are some of the things that we do. There are other responsibilities as well, but those are the big three. And then you asked about some of the challenges, and I think you're alluding to something I said when we met at the conference in Minneapolis, which is really the darkest moment for me of a job I love and a job that has been fantastic. And this is an outlier. This is by no means typical or regular. But the biggest challenge for me personally was around the time of the January 6th attack on the US Capitol, where there were similar sort of dark and veiled threats happening, not just in Minnesota, but in many states. And I got a call at one point from law enforcement in Minnesota saying they had good intelligence that there were some armed protesters, armed that might come to my home. And they suggested that maybe my wife and my kids and I ought to get out for a little while, go to a hotel or go to some friends or something like that. Now we decided to stay and nobody ever did come. I want to make that very clear. But that was a scary time. And if it were just my wife and I and armed protesters have come to our home, we might just draw the shades and watch Netflix or something like that. But we have two little kids. So it wasn't just about us and our comfort level. And we had to think about that. And that's not something anyone signs up for when they run for an office or take any job, elected or non elected. And you know, again, that was the worst moment. It hasn't happened since then. Nothing bad did happen that day or during that time period. But that was the biggest shock and the one that was the biggest challenge when it started to hit home, literally for me and for my family. But overall, it's amazing work. It's gratifying work, it's fulfilling work, it's, I hope, important. Work. And I wouldn't have it any other way.
[35:52]
Liberty Vittert
I know. I have a final question. We ask every one of our guests a magic wand question. So for you, if you could wave your magic wand and have every voter do one thing this election before voting, voting, what would it be?
[36:07]
Steve Simon
Oh, okay. So magic wand. And every voter does one thing before they cast their ballot.
[36:13]
Liberty Vittert
Yes, sir.
[36:14]
Steve Simon
Great question. I would say, honestly, given the climate, the sort of overall political climate, I would say, somehow in your head, demonstrate some political empathy. By that, I mean, no matter how passionate you are about a particular candidate at any level, not just the one everyone talks about, President of the United States, just remember, there are many, many people who are voting differently from you. And don't assume bad motives about the people who are voting completely opposite of you. You may think they're dead wrong, but try to resist the temptation to believe the worst about someone. It's okay to believe that they're just wrong. They probably think you're wrong, too. But just try to have a flash of empathy about the fact that in the case of a national election, millions upon millions upon millions of people are voting in a way different from you, and just understand that that's the case. Even if you believe they're wrong, and just try to reconcile that. Don't believe the worst about someone who's voting different from you.
[37:25]
Shao Meng
Well, that's a terrific advice. I think these days, that is one biggest problem, Right? We do not assume people have the positive intention. That's always a problem. And I remember at the workshop or the conference, you said that you're hoping this election will be a high turnout and a low drama. Do I quote you correctly? You're right. And I thought that was just a very, very good, you know, advice that we all need to vote, but we should really try to vote as, you know, peacefully as possible. Thank you, Secretary Simon, for your time, and most important, for what you do. Thank you.
[38:05]
Liberty Vittert
Thank you.
[38:06]
Steve Simon
Thank you. And I really appreciate the opportunity and I want to thank and congratulate HDSR for its work, for its longevity, and I hope for many new and exciting projects to. And I'm at your disposal. Come back in a few months and hopefully we can collaborate again.
[38:24]
Shao Meng
Thank you. We really appreciate that.
[38:28]
Liberty Vittert
Thank you so much for listening to this week's episode of the Harvard Data Science Review podcast. To stay updated with all things HDSR, you can visit our website at HDSR, MITPress, MIT.edu, or follow us on Twitter and InstagramHDSR. A special thank you to our executive producer, Rebecca McLeod, and producers Tina Toby Mack and Arianwen Frank. If you liked this episode, please leave us a review on Spotify, Apple or wherever you get your podcasts. This has been the Harvard Data Science Review. Everything Data Science and Data Science for everyone.

Summary5 min read

Harvard Data Science Review Podcast: "AI and Elections: A Conversation with Secretary Steve Simon of Minnesota"

Release Date: June 28, 2024
Host: Liberty Vittert
Co-Host: Shao Meng
Guest: Steve Simon, Secretary of State of Minnesota

Introduction

In this compelling episode of the Harvard Data Science Review Podcast, hosts Liberty Vittert and Shao Meng engage in an in-depth conversation with Steve Simon, the Minnesota Secretary of State. As the nation gears up for the 2024 presidential, Senate, and House elections, the discussion centers on the intersection of artificial intelligence (AI) and electoral processes. The conversation delves into election security, the role of AI in amplifying disinformation, and the measures Minnesota is implementing to ensure fair and accurate elections.

Role of the Secretary of State in Elections

Steve Simon begins by clarifying the role of the Secretary of State's office in Minnesota's election system. He emphasizes the decentralized nature of vote counting, stating:

"We do not count votes. And that is the case in almost every other state. The Secretary of State does not actually count or tabulate votes." (02:04)

Simon explains that the Secretary of State’s responsibilities include knitting together various elements of the election system, providing legal and policy guidance, certifying election equipment, managing critical databases, and engaging in policy work with legislative bodies and federal partners. This delineation ensures a robust and transparent electoral framework.

The Emergence of AI in Election Security

As AI becomes a more significant factor in elections, Simon addresses its dual-edged impact. He notes:

"I do view it [AI] as a new way to amplify and magnify existing threats. The existing threat is, for example, election disinformation... this is the first election... when those considerations have been on the table." (04:14)

Simon highlights that AI, particularly through tools like deepfakes, can enhance the spread of disinformation, making it a pivotal concern for election officials. This marks a new challenge distinct from past election cycles, necessitating innovative approaches to safeguard electoral integrity.

Combating Disinformation

The conversation underscores the importance of media literacy in combating AI-driven disinformation. Simon explains the office’s strategy to promote trustworthy information sources:

"We're going to continue to urge people... to really seek out trusted sources on that issue. In other words, don't just trust what's in your social media feed... go to a trusted source." (06:00)

By encouraging voters to verify information through official channels, the Secretary of State aims to mitigate the impact of false narratives surrounding election mechanics, thereby fostering a more informed electorate.

Education and Media Literacy

Simon advocates for the role of the education system in enhancing public understanding of election processes. He shares Minnesota’s initiative to incorporate civics education in high schools:

"People would be less susceptible to some forms of disinformation if they understood that very basic piece of the architecture." (09:41)

By teaching students the fundamentals of government structure and election administration, the state aims to build resilience against misinformation from a young age, fostering long-term benefits for democratic participation.

Election System Security and Transparency

A significant portion of the discussion focuses on ensuring the security and transparency of election systems. Simon details the public accuracy tests conducted for election equipment:

"Anyone off the street can come. They have to trot out that elections equipment in a pre-noticed public meeting called a public accuracy test." (14:06)

These tests involve stress-testing voting machines with various attempts to disrupt their functionality, ensuring they perform accurately under real-world conditions. Post-election audits further reinforce this transparency by reconciling reported results with paper ballots, providing multiple layers of verification.

Collaboration and Coordination Efforts

Simon emphasizes the importance of collaborative efforts both within the state and across federal agencies to secure elections:

"It's essential that we reach out to partners... we have to partner with others. We have to go to those who can tell the difference." (06:00)

He highlights Minnesota’s participation in the National Association of Secretaries of State (NASS) and partnerships with the Department of Homeland Security and CISA, which facilitate information sharing and best practices to enhance election security.

Initiatives and Future Plans

Discussing future initiatives, Simon mentions a tabletop exercise conducted to explore AI-infused scenarios affecting election administration:

"We convened a group to do a tabletop exercise at a military facility... and we had a lot of intense, lengthy discussions over exactly that." (26:06)

Such exercises aim to anticipate and prepare for potential AI-related threats, ensuring that Minnesota remains proactive in addressing emerging challenges in election security.

Personal Insights and Challenges

Simon shares a personal anecdote highlighting the pressures faced by election officials in the wake of political unrest:

"The darkest moment for me of a job I love... was around the time of the January 6th attack on the US Capitol." (31:10)

He recounts receiving threats and the subsequent need to protect his family, illustrating the risks and emotional toll that come with overseeing election integrity in a polarized environment.

Final Thoughts and Advice

In a heartfelt conclusion, Simon offers advice to voters to practice political empathy:

"Just try to have a flash of empathy... Don't believe the worst about someone who's voting different from you." (36:14)

He urges voters to recognize the diverse perspectives within the electorate and to approach political differences with understanding rather than suspicion, fostering a more respectful and cohesive democratic process.

Conclusion

This episode of the Harvard Data Science Review Podcast offers a comprehensive exploration of the intricate relationship between AI and elections. Through Steve Simon’s insights, listeners gain a deeper understanding of the challenges and strategies involved in maintaining election security and integrity in the digital age. The conversation underscores the importance of transparency, collaboration, and education in safeguarding democracy, making it a valuable resource for anyone interested in the nexus of data science and electoral processes.

Notable Quotes

Steve Simon on the Secretary of State's role:

"We do not count votes... We compile and aggregate others counts, but we don't do the counting." (02:04)
On AI as an amplifier of threats:

"I view AI as a new way to amplify and magnify existing threats." (04:14)
On media literacy:

"Seek out trusted sources... go to a trusted source." (06:00)
On public accuracy tests:

"Anyone off the street can come... to make sure[...] the machines are accurately tabulating." (14:06)
Advice to voters:

"Have a flash of empathy... don't believe the worst about someone who's voting different from you." (36:14)

This detailed summary encapsulates the key discussions, insights, and conclusions from the podcast episode, providing a clear and engaging overview for those unfamiliar with the original conversation.

Loading summary

Transcript36 lines

[00:02]
Liberty Vittert
Welcome to the Harvard Data Science Review Podcast. I'm Liberty Vittert, the feature editor of the Harvard Data Science Review, and joining me is my co host and Editor in Chief, Shao meng. After the 2020 United States presidential election, there were widespread claims that the election was unfair, insecure, and in some instances, outright stolen. This outcry put election officials under intense scrutiny, particularly those in the Secretary of State offices across the country. As we approach another presidential election season, it's crucial to address the most significant election safety concerns for the 2024 presidential, Senate and House races. Today we delve into these issues with Minnesota Secretary of State Steve Simon. How does election reporting function? How will AI influence this election? Is your vote really being counted? This is a special edition of the podcast as Secretary Simon joined us for a Conversation with Leaders piece earlier this month. Please join us for these discussions and more on the Harvard Data Science Review podcast.
[01:11]
Shao Meng
Well, thank you so much again, Secretary Simon, for accepting my invitation to talk to hdsr. Conversation with Leaders I really enjoyed your presentation in May at the University of Minnesota's Data Science Initial Conference as a statistician and a citizen, because I found your presentation both insightful and informative, just like how you started in your presentation there. Could you share with HDSR's readers and listeners what your office does or does not do regarding the election? And are these things typical for other states or there are considerable variations among them?
[01:45]
Steve Simon
Well, first of all, I want to thank you for having me. Thanks for initiating this discussion. I really appreciate your interest in the topic and I really appreciate the work of hdsr, what I've come to know and understand of it, and congratulations on your recent milestone and I wish you the best. So thank you for having me.
[02:04]
Shao Meng
Thank you, thank you.
[02:05]
Steve Simon
I mean, as Secretary of State in Minnesota, as is the case in almost every state, we oversee the Minnesota election system. And it's that way, as I say, in most states. So here's what that means. I'll start with what it doesn't mean. We do not count votes. And that is the case in almost every other state. The Secretary of State does not actually count or tabulate votes. That happens at the local level in towns, in cities, in counties. It's very dispersed, it's very decentralized. That is work that is done at the local level by local people, often the friends and neighbors of the voters themselves. We compile and aggregate others counts, but we don't do the counting. Similarly, we don't hire or train or pay the actual election workers. In Minnesota, we call those election judges, but they can be called poll workers or election workers. That too is a local function and responsibility. But here's what we do in our office. We knit together those various elements of the system. We provide legal guidance and policy guidance. We certify all the elections equipment, the tabulating equipment for use in Minnesota. We design and run the giant databases that are required by federal law that are sort of the spine of the system. And maybe most importantly, we do the policy work at the Minnesota Legislature and with our federal partners in Washington, D.C. so as a result of all of those responsibilities, I like to say that I and we are in the democracy business, and it is quite a time to be in the democracy business. But I think understanding that setup and that division of labor and division of responsibility is really important for understanding the system itself.
[03:51]
Liberty Vittert
Secretary Simon, I can't help but feel like this is a significant election year and especially with sort of major players and what feels like a new external factor of AI in the mix. Do you see any major shifts from your perspective as Secretary of State? Is there anything you're really looking at?
[04:14]
Steve Simon
Yes, and I think it was embedded in your question. There's no question in my mind that AI related considerations are pretty new in 2020. We were talking about a lot of things. We were talking about the challenges of COVID and a once in a century pandemic. We were talking about election disinformation, as we almost always do. But this is really the first time in a sustained way that we who do this work, we who are in the democracy business, not just secretaries of state, but others in every state who do and oversee and administer elections. This is the first time we've really been focused on the AI issue. And as I like to say, I don't view AI as a particularly new threat in and of itself. I do view it as a new way to amplify and magnify existing threats. The existing threat is, for example, election disinformation. That problem isn't new, it's very old, actually. But this is a means of amplifying and magnifying that problem, that threat, which we haven't discussed now at hdsr, you may have been discussing it for a lot of years, but to lay people to people not immersed in this area, this is the first election, certainly the first presidential election, when those considerations have been on the table. So that feels new, it feels different, it feels bigger, it feels like a new kind of challenge. So I would really point my finger there as the thing that most distinguishes this election cycle from the Ones that I participated in in the recent past.
[05:52]
Shao Meng
So in what ways are you seeing AI's impact on the election? On the ground level, what should the average voters be concerned with?
[06:00]
Steve Simon
Well, it's a matter of how we all take in and react to information of all kinds. And in this respect, as I'm sure you know, given your expertise, the elections area is not so different from any other area. AI and deepfakes in particular can impact the way we perceive information in a hundred different areas of our lives. Election is just one, and it's a timely one, and it's the one that I do for a living. So I don't pretend that this is necessarily unique, but on the ground, it means a couple of things to me anyway. One is it emphasizes the need for close collaboration. We're one office among many that touches this area of our life, this area of democracy. And it's essential that we reach out to partners on this, because so many people have different areas of expertise. I'll give you an example. When it comes to AI, we in the Office of Secretary of State, we aren't the technical experts, we aren't the data experts really, in the sense that when it comes to detection, is this thing, is this phone call, is this video, is this audio clip, is it or is it not a fake in the sense that it was done by generative AI, and it isn't what it purports to depict? Right? We aren't the detection experts. We really leave that to others. But that means we have to partner with others. We have to go to those who can tell the difference. Now, once it's detected, we do have certain areas of responsibilities. It's we who can speak to voters. It's we who can use the channels we already occupy to make sure voters are getting good and accurate information. But to me, on the ground, that's one of the things it looks like. Now, that's not necessarily, as you ask, what voters would see. Right? That's what we who administer elections would see. What voters will see from our office, and not just our office, from many around the country and around the state, is more and more attention paid to media literacy, urging voters to show a healthy skepticism about the messages they receive about the mechanics of elections. I'm not talking about who they should vote for or which political party or candidate. We're not in that business. But in terms of the rules of elections, the who, what, why, when and where, we are going to continue to urge people, especially in light of AI, to really seek out trusted sources on that issue. In other words, don't just trust what's in your social media feed, at least not alone and by itself when it comes to those rules, go to a trusted source. We hope that includes our office. For those who live in Minnesota, they'll go to our website. But if it doesn't, for whatever reason, go to one trusted source. That is something that has integrity to you. It could be a city, it could be a county, it could be someone else. But really, practitioners, people who are administering elections, before you believe something in your Facebook feed or that you see on X or elsewhere in the social media world, go to those trusted sources. So on the ground, what I think in this pre election period, as we're approaching November, it's more and more emphasis on that voters will see invitations for them to be more discerning and discriminating. We want people to have a healthy skepticism, but not an unhealthy skepticism. And where that line is is a difficult one. But we want people to view assertions about the election system carefully.
[09:33]
Shao Meng
Let me follow up quickly how much the education system can help in this process, like universities, even high schools.
[09:42]
Steve Simon
I think it can help a great deal. And in Minnesota, we recently had a debate about that. There had been a long running discussion, finally resolved last year in 2023, about establishing some sort of civics requirement in high school in Minnesota. There was none before and there is one now. Now, civics can mean a lot of things, but what it means at a fundamental level is just very basic things about teaching young people the architecture of government. Who does what, what does a court do, what does the executive branch do? What does a legislative branch do? And maybe on a deeper level, what do people like a Secretary of State do? I mentioned at the outset, in response to your question, one of the things we don't do is count votes, right? People would be less susceptible to some forms of disinformation if they understood that very basic piece of the architecture. And what I'm referring to is the fact that if any of your listeners would do a quick 10 or 30 second Google search, they would quickly see that there are people out there asserting falsely and wrongly that I, and not just me, but my colleagues around the country who have similar portfolios and roles are secretly scheming in this pre election period in some back room with our secret algorithms. And we are plotting and scheming, so they say, falsely, to spike the vote totals for the candidates we like and to depress the vote totals for the ones we don't like. Now that Just isn't a matter of opinion. Someone who understands the basic architecture would then understand, hey, Secretaries of State do a lot of things, and I like some of those things and don't like some of those things, but I know that they don't count votes. That's one thing they don't do. So this thing must be false. That's one very basic example of the way in which some education, particularly of young people, can be helpful and can pay dividends in the long run. People understand who does what in our country, just mechanically speaking, they are less susceptible to these false appeals on social media or anywhere else.
[11:48]
Liberty Vittert
I think that that makes me bring up the question of whether AI, you know, not necessarily disinformation, but could it be affecting the election process at a systems level? You know, that was a huge part of the discussion in 2020, the Dominion Systems, you know, all this kind of stuff. Could AI influence the way elections are run? Or really the security, I suppose, of election data? Or is that also not true?
[12:14]
Steve Simon
Well, as we're coming to discover, at least those of us who are relatively new to this area, you know, AI holds a lot of promise and possibility in many ways. It can have a positive impact and influence. It can make life and processes across the board more efficient and more effective and more surgical and. And all the rest. To that extent, it can be a real benefit. But from a systems perspective, could it introduce instability or mischief or something like that? I suppose it could, which leads to a renewed emphasis in our case anyway, on cybersecurity, for example, a really important topic just to make sure our systems are secure, whether it's the databases we run or the transmissions between us and counties and cities, to make sure there's no level of mischief there. So the good news is I haven't seen an example yet of that kind of mischief or instability, but as I'm sure you and your listeners know, it's always a question of hoping for the best, but planning for the worst. And that really is the name of the game this election season. So whether it's cybersecurity or other means, we're always looking to make sure that from a systems standpoint, things are secure. And sometimes that just means using your imagination to try to predict or figure out what could go wrong. And we're doing some of that with a lot of partners at the federal level and at the state level.
[13:50]
Shao Meng
Thank you, Secretary Simon. Now, minasuda is often seen as a golden standard for election data reporting and accessibility. How does the office ensure the elections are safe, accurate and Efficient, particularly in terms of reporting and other large scale election tasks.
[14:06]
Steve Simon
Right. Well, I mean, part of it is to have a redundancy in place. Part of it is to make sure these things are tested, tested often, tested frequently, tested regularly. And then there's another public accountability piece as well. And I think sometimes this gets overlooked in these discussions, which is to say it's not just about the systems being effective and safe, it's about the perception that they are as well. And what that means to me anyway, and to many of my colleagues around the country, I think they feel the same way, regardless of their political orientation or anything else, is transparency is making sure the public understands what some of these processes are in terms of security. Let me just give you an adjacent sort of example from Minnesota. And I think other states, not everyone, but several, have this same thing. So take the issue of elections equipment, the election tabulators. So there are some that assert wrongly that these tabulators have been, whether in Minnesota or elsewhere, switching votes from candidate A to candidate B. Now, speaking only for Minnesota, one of the reasons that I know that's not true is all the work we do leading up to the election and after the election in looking at the performance of this equipment, one of the things we do before the election in terms of security is we have something in state law called a public accuracy test. Now I know before everyone's eyes or ears glaze over, that sounds like a perfectly bureaucratic exercise, a public accuracy test. Or Pat, I get how that sounds, but here's what it really means. It means under Minnesota law that every jurisdiction, whether it's a county or a city or a town that owns and uses tabulating or elections equipment, they must, within a couple of weeks of every election, not just a general election, a primary, a special election, every election, they have to trot out that elections equipment in a pre noticed public meeting called a public accuracy test. Anyone off the street can come. There are no engraved invitations. You're not limited to the one that's in your area or where you live. You could go to one three hours away if you wanted to. And it's a pre notice public meeting where anyone off the street can come in and watch the elections. People basically try to trick the machines. That's what these tests are. They go in there, they'll put stray marks, they'll bend the paper, they'll circle or underline instead of filling in the oval. They'll overvote, they'll undervote. The point is to try to trick the equipment and only if that equipment performs perfectly, will it be allowed for use in the subsequent election? So that's part of how we sort of stress test these systems and in this case particular equipment. And I think that's been successful. The more people can learn and know that this even exists, they don't have to attend. People have lives and families and jobs and all the rest. They're not necessarily going to go to these. But knowing that they exist, I think can give people peace of mind. And beyond just peace of mind, it actually does improve the performance of the equipment. It's public, but even if it were private, it's a way that you're testing the equipment beforehand to make sure it works.
[17:19]
Shao Meng
Quick follow up question that you said. You know, one thing you need to building duplication duties test. I assume you need to have a lot of people to work on those things. Now we have a lot of listeners, data scientists or AI specialists or whatever they are. And just for their benefit, suppose some of them wants to work for your office or generally for government agencies. What kind of skill sets do they need? In addition to the technical side, what are the other things they should learn to equip themselves in order to contribute to this great process?
[17:49]
Steve Simon
Yes, that's a great question. I don't think I've ever had that question before. That's a really good one. I would say a number of things. One is a belief in the work. As corny as that sounds, I think that that really does count for something. The belief in being in the democracy business. That is important work, it's worthy work. Another one though is in terms of skills. They're the technical data skills, of course, which always apply and are useful and important, but also collaboration and cooperation, not just internally in our office, where we're always trying to build a better mousetrap, so to speak. We're always trying at least to be in a process of continuous self improvement. But also cooperation outside the walls of the office, with other entities, with other partners, both in the nonprofit sector and in the public sector, other layers and levels of government as well. Those are some of the things that we look for. Again, as I said, in this business, as in so many others, we're not unique. It's a question of hoping for the best, planning for the worst, and being creative about what that really means. So it's that creativity, that sense of collaboration. Those are the sort of, oh, you might say softer, intangible sort of skills, in addition to degrees and particular technical expertise.
[19:08]
Liberty Vittert
Secretary Simon, I've heard, you know, different states have different systems and that you're speaking obviously from Minnesota, but do you think that there's a chance the existing confusion from these varying state election systems that AI could make this issue worse for voters? Is there a reason that states should start coordinating to sort of create more uniform election rules and regulations, or is there a really good reason for having different systems in different states?
[19:37]
Steve Simon
That is such a well timed question because it's not only come up recently, it comes up constantly in the sense that we're always trying to figure out what the best division of labor and division of power is. Now, I will say that the rough consensus in this country, with some exceptions, is that the time, place and manner of elections, the default position is up to the states, with huge exceptions. I mean, I'm sure most of your listeners are familiar with the epic and important 1965 Voting Rights act, which really stepped into this space where in light of what was happening in certain parts of our country, Congress stepped in. It upended that longstanding consensus, at least on a somewhat limited basis, and said, yes, our default position is that the 50 states have autonomy in election administration. But in certain extreme cases like this, we're going to step in and we're going to impose a federal rule. So from now on, no, you are no longer allowed to have poll taxes. No, you are no longer allowed to have literacy tests for voting, or so called character tests for voting. We know how that went down in the 60s in certain states in certain parts of our country. So there Congress stepped in and said, I know that's our default rule, but we're imposing ourselves here. Even that is more the exception than the rule. We remain a country of basically 50 plus individual laboratories with some federal oversight and some federal rules, like the Voting Rights Act. You used an interesting word, and maybe I'm reading too far into it, but I give you credit for using it, which is the word coordination, meaning can states coordinate without necessarily Congress mandating something? Because I think that would run into a lot of state opposition, because states tend to want to have that autonomy and that independence and that authority. But coordination implies to me anyway something different, something a little more voluntary, something like finding a consensus that all states can adhere to or move towards. And I think we've seen that in some areas, it may not be in every area, it may not be in an area like voter registration. But when it comes to some of the technical matters that we've been discussing today, I don't see why we can't and shouldn't move towards closer coordination as you put it, which isn't necessarily Congress stepping in and mandating things, although I support some of that too. A good measure of that. But even short of that, states talking more, collaborating more, trying to harmonize their approaches more, particularly in areas of data. Not so much rules around, like I said, voter registration, but data, data used, data disclosure, security protocols and practices. Those are things that don't necessarily require an act of Congress and a one size fits all. But more and more states can learn from one another and maybe over time it won't happen immediately move towards some rough standardization on certain issues like these issues.
[22:48]
Shao Meng
Thank you. Secretary Simon, Besides coordinating or collaborating with other states, what other organization does the Secretary of State's office work with to ensure our elections are secure, fair and well reported?
[23:01]
Steve Simon
Well, the first and foremost organization that I have to mention, just because of the place of respect that it occupies, is the national association of Secretaries of State, or nass. And if I can say, it's the oldest such organization in America, the oldest organization, bipartisan organization of statewide elected officials, founded in 1904. And it is what it sounds like, it's the national association of all the Secretaries. We get together physically twice every year, once in the winter in Washington, D.C. and once in the summer on a rotating basis in various locations across the country. And we talk and we compare notes and we exchange best practices. And even when we're not meeting physically, we have one off or occasional zoom meetings on particular topics where we, and not just we, but members of our staff will collaborate and trade and try to learn from one another. So I can't say enough in praise of NASS as just a great connector between and among Secretary of State offices. But there are other organizations as well that we deal with. We have a great partnership, I think, on security and data issues with the federal government, with the Department of Homeland Security in particular, and even on a more granular level with cisa, which is of course part of the Department of Homeland Security. Just for the benefit of your listeners, we in the elections area took a big step forward in 2017. 2017 was the year when the Department of Homeland Security designated our election systems as critical infrastructure. And some of your listeners are maybe aware of what that means. Before then, I hadn't been familiar with that term, but critical infrastructure was a post 911 sort of department of Homeland Security designation that applies to various segments or sectors of national life. Hospitals, military bases, the power grid, the banks. You get the idea. Key core sectors of our national life that deserve special status and special government help and special Protections and defenses. And in 2017, this became, I believe, the 18th sector of our national life to be given that designation. What that has meant is a locked in, really effective, I think, partnership with a lot of federal entities, some of which I just mentioned. So they've been a great partner as well when it comes to security and data and technology and frankly, intelligence. One of the advantages of having this critical infrastructure designation is that I and a small number of folks in my office, and this is the same in other Secretary of State offices, we get intelligence briefing and security clearances that enable us to do that. So they've also been a great partner as well on all these issues.
[25:52]
Shao Meng
Thank you. It's great that you work with all these different agencies. I want to also ask you about any new initiatives that your office is doing, maybe with some of these partners in the AI space.
[26:06]
Steve Simon
Yeah, well, in terms of AI initiatives, I mean, one of the things, I don't know if this qualifies as necessarily an initiative, but one thing that we started out doing at the beginning of this year back in January, is we convened a group to do a tabletop exercise at a military facility in Minnesota. And we brought in the Department of Homeless Security in cisa, FBI, the US Postal Inspection Service, many state Minnesota law enforcement organizations, and critically, a majority of our Minnesota counties were represented as well. And what we did during this tabletop exercise is we gamed out a lot of scenarios that could affect election administration, including, including very much AI infused scenarios. Now, I'll spare you all the details, I can't give a roadmap to the bad guys, obviously, but I will tell you that we had a lot of intense, lengthy discussions over exactly that. Who does what, what is the division of labor. And so that's a new initiative. It's not so much technical in nature, but it is a level and layer of collaboration that is relatively new to do it, to be focused on this particular issue so that we can all understand who's doing what. That's one initiative that we're undertaking, you know, another initiative, again, it's not so technical in nature is what I discussed a little bit earlier, which is really intensifying our efforts to engage in media literacy to make sure that people are better positioned to take in real information and to reject false information about how our election system works, what it is, and we think that's very important as well.
[27:52]
Liberty Vittert
Is there a way that a citizen, if they are concerned about election security, can see for themselves? Are there parts of the election process that they can go watch or what would be the sort of the options there?
[28:05]
Steve Simon
Yeah, I'm really glad you asked that because I happen to believe that one great way to push back against disinformation or really stop it before it starts is to be as transparent as possible. So earlier I mentioned the public accuracy test as one way before the election in one specific area, that is the integrity of the elections equipment or the tabulating machines. That's one way that we invite scrutiny, we invite public view and participation, and we invite people to ask tough questions and make up their own minds. But maybe I should talk a little bit about what happens after the election. Because it's after the election, not before, of course, that you get allegations of the kind that we've seen in recent years, false allegations, saying, for example, that the equipment has transferred votes from candidate A to candidate B. What we do in Minnesota and that many other states do after the election is we have post election audits or reviews where those kinds of issues are addressed, where we do a formal audit and review of the system, where we reconcile the reported results with a randomly selected precincts all over the state with the paper ballots to make sure that the machines were accurately tabulating and reporting what the returns were. And even then it goes through two layers of review. We have county level review and then statewide level review. And after that, if a particular citizen or person is dissatisfied or suspicious or skeptical or even hostile, then they can go to court and marshal the evidence and try to persuade a judge of their point of view. And that judge has the authority to reopen part or all of a particular election contest. So there are multiple areas or avenues of accountability and review and checking the work. That's what happens after the election. It's not just about before, it's after as well. So we hope overall that that persuades people who are suspicious or skeptical or even hostile that there are multiple eyes on multiple people at multiple levels and that ought to give them confidence in the overall system. That's not to say mistakes aren't uncovered occasionally. But overall, the integrity of the system is good in part because there are all these avenues of accountability. So that's something that I and many of my colleagues across the country have been emphasizing in their states as well.
[30:22]
Shao Meng
It's great to hear that. Obviously, transparency is incredibly important. And Secretary Simon, that we have been talking so far is on the AI in the elections that I hope you don't mind that I'm going to ask you a little bit more personal question. Part of the reason that we run this conversation with leaders Is obviously want to inform the community, the data science community, how leaders are using data, so on and so forth. But we also like to inspire some people like to be a leader some days. So obviously, as a Secretary of State, you do a lot more than just election. So first, if you can tell our audience thinking generally, what do you do in terms of your overall charge and also what are the excitement and the challenges? Just give us the general sense. What is it like to be a Secretary of state?
[31:11]
Steve Simon
Yeah. Well, I'll tell you this. Here's what's interesting about it. So I previously served in the Minnesota Legislature. I was in the Minnesota House of representatives for 10 years. And sometimes I use a food analogy to distinguish those two jobs. Both are leadership roles, but they're a little different. Being in a legislative body anywhere in any state or at the national level is like visiting a buffet. You can load your plate up with dozens of different combinations of things. You can take five big heaps or 20 little spoonfuls or whatever, and there's so many issues you can be involved with. But in the Office of Secretary of State, the portfolio is more limited. I don't do tax policy anymore. I don't do healthcare policy or transportation policy anymore. So it's not a buffet. Rather, it's a fine meal with limited courses, four or five or more courses, delicious. You can dive into the flavors, but it's not a buffet. So what are those courses we mentioned? Elections. That's the one that gets the most attention. But we in our office and almost every other Secretary of state office around the country also has a business services department. What that means is it's the place where businesses start, and it's the place where people do other kinds of filings, all sorts of corporate and commercial related filings. So very often the very first thing that someone will do when they start a business, before they hire anyone, before they get an office, before they get insurance, the first thing they have to do is make the thing come alive legally by making it an official thing, an official entity. And you do that and much more through our office. So we like to say we're the welcome mat for Minnesota businesses. And that's really gratifying work. That's one thing we do. Another one that I'll just highlight in the top three. So elections, business services. And another thing we do is we run a program called the Safe at Home program. Safe at Home is an address confidentiality program for victims of domestic violence or stalking or sexual assault. And participants in the Safe at Home program get to have their home address, their true physical home address, kept completely secret and private for all purposes, whether it's interacting with a public entity or a private entity, like a bank or a retailer or anyone else. So the only one who knows where the participant lives is the participant, our office, and whoever the participant chooses to tell. And so we perform a lot of services for them, from getting them to vote safely to handling their mail in a safe way that's not going to reveal where they actually physically live and so on. And it's gratifying and really life saving work that I'm. I'm so really privileged to be a part of. And it has saved lives, there's no question about that. So those are some of the things that we do. There are other responsibilities as well, but those are the big three. And then you asked about some of the challenges, and I think you're alluding to something I said when we met at the conference in Minneapolis, which is really the darkest moment for me of a job I love and a job that has been fantastic. And this is an outlier. This is by no means typical or regular. But the biggest challenge for me personally was around the time of the January 6th attack on the US Capitol, where there were similar sort of dark and veiled threats happening, not just in Minnesota, but in many states. And I got a call at one point from law enforcement in Minnesota saying they had good intelligence that there were some armed protesters, armed that might come to my home. And they suggested that maybe my wife and my kids and I ought to get out for a little while, go to a hotel or go to some friends or something like that. Now we decided to stay and nobody ever did come. I want to make that very clear. But that was a scary time. And if it were just my wife and I and armed protesters have come to our home, we might just draw the shades and watch Netflix or something like that. But we have two little kids. So it wasn't just about us and our comfort level. And we had to think about that. And that's not something anyone signs up for when they run for an office or take any job, elected or non elected. And you know, again, that was the worst moment. It hasn't happened since then. Nothing bad did happen that day or during that time period. But that was the biggest shock and the one that was the biggest challenge when it started to hit home, literally for me and for my family. But overall, it's amazing work. It's gratifying work, it's fulfilling work, it's, I hope, important. Work. And I wouldn't have it any other way.
[35:52]
Liberty Vittert
I know. I have a final question. We ask every one of our guests a magic wand question. So for you, if you could wave your magic wand and have every voter do one thing this election before voting, voting, what would it be?
[36:07]
Steve Simon
Oh, okay. So magic wand. And every voter does one thing before they cast their ballot.
[36:13]
Liberty Vittert
Yes, sir.
[36:14]
Steve Simon
Great question. I would say, honestly, given the climate, the sort of overall political climate, I would say, somehow in your head, demonstrate some political empathy. By that, I mean, no matter how passionate you are about a particular candidate at any level, not just the one everyone talks about, President of the United States, just remember, there are many, many people who are voting differently from you. And don't assume bad motives about the people who are voting completely opposite of you. You may think they're dead wrong, but try to resist the temptation to believe the worst about someone. It's okay to believe that they're just wrong. They probably think you're wrong, too. But just try to have a flash of empathy about the fact that in the case of a national election, millions upon millions upon millions of people are voting in a way different from you, and just understand that that's the case. Even if you believe they're wrong, and just try to reconcile that. Don't believe the worst about someone who's voting different from you.
[37:25]
Shao Meng
Well, that's a terrific advice. I think these days, that is one biggest problem, Right? We do not assume people have the positive intention. That's always a problem. And I remember at the workshop or the conference, you said that you're hoping this election will be a high turnout and a low drama. Do I quote you correctly? You're right. And I thought that was just a very, very good, you know, advice that we all need to vote, but we should really try to vote as, you know, peacefully as possible. Thank you, Secretary Simon, for your time, and most important, for what you do. Thank you.
[38:05]
Liberty Vittert
Thank you.
[38:06]
Steve Simon
Thank you. And I really appreciate the opportunity and I want to thank and congratulate HDSR for its work, for its longevity, and I hope for many new and exciting projects to. And I'm at your disposal. Come back in a few months and hopefully we can collaborate again.
[38:24]
Shao Meng
Thank you. We really appreciate that.
[38:28]
Liberty Vittert
Thank you so much for listening to this week's episode of the Harvard Data Science Review podcast. To stay updated with all things HDSR, you can visit our website at HDSR, MITPress, MIT.edu, or follow us on Twitter and InstagramHDSR. A special thank you to our executive producer, Rebecca McLeod, and producers Tina Toby Mack and Arianwen Frank. If you liked this episode, please leave us a review on Spotify, Apple or wherever you get your podcasts. This has been the Harvard Data Science Review. Everything Data Science and Data Science for everyone.