How Americans’ Feelings Towards AI Are Getting More Complicated—and the Revolt That’s Simmering | Behind the Numbers

Summary

Behind the Numbers: An EMARKETER Podcast Summary

Episode Title: How Americans’ Feelings Towards AI Are Getting More Complicated—and the Revolt That’s Simmering
Release Date: June 9, 2025

Behind the Numbers by EMARKETER delves into the evolving landscape of digital media, offering marketers, retailers, and advertisers insightful analyses on critical topics. In this episode, host Marcus engages with analysts Grace Harmon and Henry Poudley to explore the complex emotions Americans hold towards Artificial Intelligence (AI) and the emerging tensions between AI experts and the general public.

---

1. Introduction to AI Sentiments

The episode kicks off with a brief discussion on the historical context of longstanding companies, serving as an icebreaker before transitioning into the main topic: Americans' feelings about AI. Marcus introduces Grace Harmon, a technology and AI analyst, and Henry Poudley, SVP of media, content, and strategy, setting the stage for an in-depth conversation.

2. Shifting Emotions Towards AI

Grace Harmon describes the public sentiment as a blend of excitement and caution. She states, “There is a lot of interest, especially on the enterprise side and a lot of excitement there. But I think that people are starting to see some of the more personal impacts like on jobs, on human connection that AI can have” ([05:23]).

Henry Poudley echoes this sentiment, highlighting the duality of growing enthusiasm and rising concerns. He notes, “On one side you've got genuine enthusiasm growing... but at the same time you've got genuine concern that grows around AI” ([04:02]).

3. Public Concerns About AI

The discussion delves into specific areas where Americans express apprehension towards AI:

• Deepfakes and Privacy: Many are worried about the misuse of AI in creating deceptive content.
• Political Propaganda: The potential for AI to influence political opinions is a significant concern.
• Job Replacement: A primary fear is that AI will displace human workers, diminishing their value and creativity.

Marcus references a YouGov survey, revealing that 54% of respondents feel cautious and 47% are concerned about AI ([05:57]).

4. AI’s Impact on the Workforce

A pivotal point of discussion centers on how AI affects employment. According to a Pew Research study cited by Marcus, 52% of workers feel more worried about AI, while 36% remain hopeful ([07:13]).

Grace Harmon adds that many employees feel their creativity is being undermined by AI tools, leading to resistance against their company's AI strategies. She explains, “There's just an enormous training gap. Employees are generally expected at most companies to figure out how to use it all on their own” ([08:10]).

5. Employee Pushback and Corporate Strategies

The conversation highlights instances where companies like Duolingo and Shopify have replaced human roles with AI, eliciting negative reactions from both employees and the public.

Henry Poudley discusses the backlash against such moves, emphasizing that public perception is largely negative when organizations reduce human positions in favor of AI. He remarks, “If it had been new jobs being created and training being offered for people who want to move into those jobs, I think that could have been received a little differently” ([12:12]).

6. Bridging the Gap Between Experts and the Public

A significant portion of the episode focuses on the disparity between AI experts and the general populace regarding their understanding and concerns about AI.

Henry Poudley observes, “Experts are gathering tons of information about how AI can evolve... and that's not generally how the public approaches the topic” ([18:32]). This gap is evident in differing levels of concern about job losses and the overall impact of AI.

7. Recommendations for Engaging with AI

Towards the end, Marcus seeks advice on how individuals, especially the skeptical, can effectively engage with AI. Drawing from Nicole Nunguin’s recommendations, two primary strategies are suggested:

1. Choose Your Bot: Select from various AI platforms like Microsoft’s Copilot, OpenAI’s ChatGPT, or Google’s Gemini based on your needs.
2. Enhance Search Queries: Provide more detailed and contextual information to AI tools to improve response quality.

Henry Poudley shares his personal experience of switching his default search engine to an AI-based one for a week, encouraging listeners to “jump in the deep end” and explore AI firsthand ([22:06]).

Grace Harmon emphasizes understanding the best use cases for different AI tools, advising users to identify when to use chatbots versus traditional search engines based on task requirements ([23:57]).

8. Conclusion and Future Outlook

The episode concludes with a reflection on the inevitable integration of AI into daily life and the necessity for both individuals and companies to adapt. Grace Harmon asserts, “I don't think there's that much of an option just to entirely sit it out” ([16:30]), indicating that AI's influence will continue to grow, and resistance may not be sustainable in the long run.

Henry Poudley underscores the importance of transparency and demonstrating AI’s positive benefits to shift public sentiment towards acceptance ([14:02]).

---

Key Takeaways:

• Mixed Emotions: Americans exhibit a complex mix of excitement and concern regarding AI.
• Employment Concerns: A significant fear is the potential for AI to replace human jobs, affecting employee morale and job security.
• Training Gaps: There is a notable lack of adequate training for employees to effectively use AI tools, leading to resistance and potential sabotage.
• Public vs. Expert Perception: Experts possess a more nuanced understanding of AI's potential and challenges, while the public remains cautious and sometimes skeptical.
• Engagement Strategies: For those looking to embrace AI, selecting appropriate tools and enhancing interaction methods can facilitate a smoother integration.
• Inevitable Integration: AI is becoming entrenched in various aspects of life and business, making adaptation essential for both individuals and organizations.

This episode underscores the necessity for transparent communication, adequate training, and strategic implementation of AI to bridge the existing gaps in perception and utilization between experts and the general public.

Transcript

[Marcus]

Are your brand campaigns as effective as they could be? If you're only getting insights when the campaign is over, then the answer is of course no. To make better campaign decisions, you need real time measurements. You need lucid measurement by sint. Discover the power of real time brand lift [e-mail] insights that's cint.com insights hey gang. It's Monday, June 9th. Grace, Henry and listeners, welcome to behind the Numbers, an E marketer video podcast made possible by synt. I'm Marcus and today we will be discussing how workers feel about AI and the biggest gaps between AI experts and the general public. Join me for that conversation. We have two people, our analyst covering everything technology and AI living in California. It's Grace Harmon.

[Grace Harmon]

Thanks for having me.

[Marcus]

Of course. Thank you for being here. We also have with us our SVP of media, content and strategy hanging out up in Maine. It's Henry Poudley.

[Henry Poudley]

Hey Marcus.

[Marcus]

Hey fellow. Today's facts what are the oldest companies in the world? So the other day I did the oldest company in America, but now I've gone and found a piece by Iman Ghosh of Visual Capitalists who wrote a piece a few years ago referencing business financing. That's where this information came from. And the oldest company in the world is in first place. Japanese construction company. Can't pronounce this but Congo Gumi Co. Ltd. Founded in Wen. When do you guys think this is the oldest company has been in business since 1200. Back further, a lot further.

[Henry Poudley]

Wow.

[Marcus]

Shockingly I know 1200 is a good guess. That's the oldest university I believe Oxford is around 1200. 1100. 1200 ish. 900 even further 578. This company's been in business 1500 odd years. Second place goes to a lot of them are 50. Second place goes to Austrian restaurant St. Peter Skifts Culinarium which got started in 803. And in third place is German winery and distillery Stauffelter Hof in 862.

[Henry Poudley]

I was wondering if there'd be a winery in there. That makes sense.

[Marcus]

Yeah. It should be illegal to say established in if you're not at least 100 years old. You know when people like OH, established in 2016. What you're saying is you've been doing this for nine years and that's not impressive especially compared to this Japanese construction company. Who can say We've been in business since the Western Roman Empire fell in 500 AD. That's amazing. Anyway, it's real topic how Americans feel about AI. So a is everywhere. At least it feels that way. But how are folks engaging with it? According to some may data from Pew Research, nearly 60% 60 of American adults viewed a web page with an AI generated summary like AI overviews. But just 13% visited a site of an AI generative tool and 10% looked up an AI related term. That's how people are using AI. But how do they feel about it? YouGov took two readings, one in December of last year, one in March of this one to try and gauge how Americans feelings on AI have changed. Before I give the results. Henry, can you fill in the blank for us? You get your take on this. Americans feelings towards AI have gotten more, what in the past couple of months or this year?

[Henry Poudley]

I think Americans feelings towards AI have gotten more complicated in the past few months. I think on one side you've got genuine enthusiasm growing. I think we know that more and more people are using tools like ChatGPT. I think they're cited at like 5 billion queries a week now. We've seen viral things like the studio Ghibli image generation or like the action figure yourself challenge that I saw a lot of people doing. And then like you just said, I think most people are encountering AI in search and seeing how that's changing the experience. But at the same time you've got genuine concern that grows around AI, around a number of situations. One, safety and creativity and whether we want to really give up so much creativity to AI tools. There's concerns about the environmental impact of AI and then of course there's concerns about jobs and how this continues to change and will further change how people work.

[Marcus]

Yeah, creativity is a good one and I'll come to that in a second. Grace, I want to get your take first and then we'll come back to that word, creativity, because that seems to be very important folks, in terms of how AI is sweeping the nation, the world. Grace, your word, if you had to describe how people's feelings are changing towards AI, what would it be?

[Grace Harmon]

I would definitely agree about excitement. I think, I would say cautious and skeptical. I think that like you said, there's a lot of interest. There is a lot of interest, especially on the enterprise side and a lot of excitement there. But I think that people are starting to see some of the more personal impacts like on jobs, on human connection that AI can have and like you said, sustainability. So I think there's a lot of concerns about its accuracy as well. So I would say I think there is more of a cautious sentiment growing despite you know, a lot of the fun. The fun aspects that AI can have, like the studio gibboot trend.

[Marcus]

You're. I mean, your take seems to be reflected in this data because that's exactly the word. YouGov survey strongest feeling towards AI in the last couple of months since December, cautious 54%. Then concerned 47%. They're both up a fraction. Then skeptical, which was also the feeling that had grown the most. Why were they concerned? Number one feeling was concerned. But why most folks were concerned. All of These shares between 50 to 60% of people, but most of all, deepfakes. Then the erosion of privacy and then political propaganda, the replacement of human jobs, and manipulation of human behavior were joined forth. Henry. But you just listed a bunch of, you know, facets of AI and it's. It's almost. I feel like too big of an umbrella to talk about all this because you do have so many way that it's being woven into people's lives that it's hard. It feels like it's hard to have an opinion on AI as an umbrella term.

[Henry Poudley]

I think that's fair. I mean, we talk about AI a lot in terms of writing and creativity and task a job tasks that many people do in their jobs. And a lot of times that's generative AI, but there's a whole realm of AI that's much more in the data and the infrastructure side that, you know, we could talk about as well.

[Marcus]

Yeah, yeah. Speaking of creativity, there's some research on how workers feel about AI. Pew found that last October, workers felt more worried 52% than hopeful 36%. But they're not miles apart. They're relatively close. But why might they feel worried? Writer had a March study looking into the main reasons why US workers were working against their company's Gen AI strategies. Working against them. Most 33% thought that AI diminished their value or creativity, to use the word that was just referenced by Henry in joint seconds with 28%. There were three other reasons. AI has too many security issues. I don't want AI to take over my job. And the company's AI tools were low quality. A quarter of workers said AI was adding to their workload. Grace, you've written recently about employees trying to sabotage their company's AI efforts. What did you find there?

[Grace Harmon]

I mean, I think for one thing, it isn't necessarily always going to be sabotage in terms of turning in fake numbers or completely falsifying information about how it's impacting your job. A lot of that might just be not giving your Employer insights on what you're learning. I think that there is not necessarily as much excitement from employees about AI as there is from executives, the people who are in the trenches actually using it or being able, able to see where it's useful, where it's not. And I think another really big issue is that there's just an enormous training gap. Employees are generally expected at most companies to figure out how to use it all on their own. A lot of times they need to pay for the tools themselves or what the employer is giving them is just not that fit for their job. You know, like Henry was saying it, it isn't always just generative AI. That's the one catch all category for AI. There's a lot of different products and a lot of different companies. And I'd say in terms of morale, there's also just a lot less trading opportunities afforded to women and older generations, which is a big issue. Yes, I think that the people who are really using are figuring out where it's actually useful and whether or not their employers are giving them the resources.

[Marcus]

Yeah, yeah, it does seem like the wild west at the moment. Henry, I'm wondering two things. One, will that change? Will it be that most companies in a few years do have a policy, do have training, or is it going to be the kind of test and learn phase for a much longer period of time? And then two, this idea about folks not really knowing how to use it kind of being left to their own devices, not really knowing how to measure success when it comes to AI, because there was a few data was saying employees thought AI chatbots sped up work, but less people thought that it improved the quality of that work. So trying to figure out when and where they should be applying AI in their jobs.

[Henry Poudley]

That's an interesting question because on one hand I do think the need for training is going to continue to grow, but I don't think it's going to be so easy for companies to decide to invest in that training. That's a resource. I think that's a resource question. And I think what you're seeing first is job changes as a result of AI. And you're hearing a lot of perspective about how the new AI jobs of the future are going to emerge. That you know, people who are displaced in jobs where AI is now pretty good at taking, taking on that task are going to move into more like an operator role or something. Right. But they need to be trained to do that. And I think that we, we need to see companies make those investments. And I'm not entirely sure that there's a lot of anecdotal evidence right now that we're seeing that.

[Marcus]

Yeah. How likely are we, do you think, to see a significant consumer or employee pushback over the next 12 months or so? Grace, does it feel like there's a brewing consumer employee revolt happening or is that just because the technology is so new and that will eventually fizzle out, or do you think it will build into something more significant?

[Grace Harmon]

I mean, yeah, absolutely. I don't know about to the level of a revolt, but to a pushback, absolutely. Like, you know, thinking about Duolingo coming out and really identifying themselves as an AI first company, how poorly that's gone over. Yeah, I'll be. I'll be interested when their next earnings report comes out to see how subscriptions have been affected because I think.

[Marcus]

Interesting.

[Grace Harmon]

It really will. So.

[Marcus]

Yeah, yeah, that's a good one. So the, the context there, it's written about by a lot of people. But I was reading about in Fortune, Mattie Merritt and Morning Brewer explained the Duolingo, as Grace was saying, CEO coming out and saying they're getting rid of contract employees and replacing them with AI. It's a language learning company. They will only allow new hires once teams prove they can't automate the work. Henry, they're not the only ones. Grace, you were saying Shopify has done this as well as well as a few others. What are your thoughts on this stance from Duolingo and the idea that this movement, if you want to call it, that these frustrations are going to build into something more.

[Henry Poudley]

And I think that's. It's natural to see that being frustrating. Right. Because the way Duolingo positioned it was a reduction. And I don't think that we should expect the public to see a reduction in a positive light. Had it been new jobs being created and training being offered for people who want to move into those jobs, I think that could have been received a little differently. And I think Klarna had a similar situation. Go on. Where they famously talked about a year ago how much they had outsourced to AI in the customer center. And I think now they're hiring humans back. So the pushback is really going to be based on whether the public see the positive or negative and what happened. And when it's just a reduction, I think it's hard for a positive to be seen.

[Grace Harmon]

Yeah, I don't think it helped anything either. The contractors are kind of more vulnerable. You know, they're not getting benefits, things like that.

[Marcus]

So, yeah, you were saying about being seen. Henry and that leads me to some, I think Grace, you had Some, some research. 80% of UK consumers think AI use in customer service should be disclosed some quality. And it sounds like a lot of people are okay with the technology. They just would like to know when it's being used, how it's being used, that it's not being used necessarily to take people's jobs directly. There's a lot of workers excited about AI. Research from Henley Business School in the UK found 56% of full time professionals, over half were optimistic about AI advancements. However, 61% said they were overwhelmed by the speed at which the technology developed. Henry how do you tackle that, how do you tackle that excitement and enthusiasm and try to make sure that the speed of development isn't muddying the water, so to speak, in terms of how people feel about AI?

[Henry Poudley]

Well, I mean I think transparency is probably the best way forward for now even if the tool itself is very much managed by human intervention. I think the public, because of this cautious feeling that folks have, wants to know and so I don't think you can go wrong being overly transparent AI use But again I think sentiment will change when the positive benefit is overwhelmingly seen. And so if people see that opportunities are being created, if there is efficiency that is gained, that efficiency is benefiting people either in their jobs or in their day to day lives. I think overall you need to demonstrate that kind of positive reaction through those results.

[Marcus]

Yeah, demonstrating the positive of signs of AI. Grace I'm wondering if there, do you think there's a, is there a time limit here? Are people just going to wait to kind of get through all of the tough parts of it and they're figuring this out and you know, eventually when AI starts to show how capable it is, a lot of the positive things it can, it can do, will people still be ready to, to welcome it with, with, with open arms? Because there's a, there's a few quotes here Mithras, that you won't BBC article from someone. There was a quote in the article from someone saying why would I bother to read something someone couldn't be bothered to write? So it does seem a bit of a sentiment of a pushback to articles being written, pieces being written, art being created, music being developed with the use of AI? And then there's this other piece of data as well. Potential employees. So is the duolingo replacing employees, contractors with AI that can do the job. So that's people at the company being pushed out. People also seem struggling to get a foot in the door. There's a recent Oxford Economics report analyzing the US employment rate, sorry, unemployment rate for those aged 22 to 27 with a Bachelor's degree. That's basically recent college grads. They looked at a three month moving average and their jobless rate of recent college grads was creeping up. Was closer to 6% in April compared to the just above 4% for the overall workforce. So what do you make of the. The potential stop clock, the countdown for which AI has to get its act together before people kind of wash their hands of it?

[Grace Harmon]

I don't think there's, yeah, I don't think there's that much of an option just to entirely sit it out. Companies are going to use AI in a lot of different facets. Just because you're not choosing to use ChatGPT or Perplexity or more direct tools like that doesn't mean you're not going to be engaging with it. I guess I'd also add about the jobs aspect. I don't know that that's all AI. There is a lot of economic factors in terms of, you know, how recent graduates are able to get jobs or if they're able to get jobs.

[Marcus]

That's a good point.

[Grace Harmon]

I don't see there being a public sentiment of we are all going to just kind of wait it out and see, you know, how this develops in a few years. I could see that happening a bit more on the enterprise side and company side, especially for people who are having to pay for those tools. You know, if you, you know, like there's been pretty low adoption of Copilot, I think that's starting to pick up. But just companies weren't seeing that there was that much of an added value from it. And that is a tool you are paying for. So I'd say the people who are paying for those tools, you know, if companies are maybe incorrectly assuming that they have created a product that's good enough to monetize, those are some of the, some of the tools that I think we might see people waiting out on, giving it a bit more time to simmer and cook and get better.

[Marcus]

Yeah, there does seem to be, I don't know, it's an education gap or two different kind of realities playing out in a sense between AI experts and then the general public. And I guess that's the case with a lot of facets of life. But in this particular instance, Henry P. Research was looking at the share of adults and AI experts, what share of them are very concerned about different aspects of AI. The biggest gap between the two parties was regarding their concerns about AI leading to job loss. There's a 26 point gap between the two parties and then the next largest gap between the two was AI leading to less human connection and then concerns about AI impersonating people. Where do you think the biggest gap exists between experts and the general public when it comes to AI attitudes and concern?

[Henry Poudley]

Yeah, to me the sentiment gap makes a lot of sense. I mean, experts are gathering tons of information about how AI can evolve over the years and how it's going to change industries. They're deeply invested in research, they're forward looking in their approach to the topic. And that's not generally how the public approaches the topic. And the public bases their opinion on what they see on the day to day. And if the news is generally bad or, you know, there's a, there's a pervading feeling of unease around the growth of AI. I think that's what's going to drive most of that sentiment.

[Marcus]

Yeah, yeah. Perception is reality. We were talking about this. Even with regards to tariffs, I think very few people would know the exact price of all of the things they buy at the grocery store. But the idea, you know, reading articles about it, the headlines, seeing the news saying prices are going up, you probably couldn't even tell if the price had gone up on specific items, but you just feel like they are. And so that starts to affect your way of thinking. There were some gaps between the two parties, but a lot of them, a lot of times they were on the same page. Actually when it came to the concerns about biases in decisions made by AI, AI spreading inaccurate information and people not understanding what AI can do, experts in the general public were actually in lockstep pretty much. I think the biggest gap might be people's awareness of how often they actually are using AI. Grace, you were kind of alluding to this Raya, that people, it's hard to just sit on the sidelines and wait out because so much of what we do already has AI built into it, baked into it. Our Gajo Sevilla was saying AI is both unavoidable online and also barely noticed, citing a March Pew research study where 93% of over 2 million page visits by a thousand adults touched a page mentioning AI. However, while 65% of users saw AI related terms in search results, just 0.05%. So basically no one of total page visits involved substantive AI engagement, basically meaning most people encounter AI incidentally. He wrote, not intentionally. They're supporting research from Gallup and Telescope as well that found nearly all Americans use products that involve AI features, but nearly two thirds don't realize it. So if AI is here and a lot of people are experimenting with it, or being asked to experiment with it by their employer or doing it on their own account, maybe, as we said at the beginning, Grace, you nailed it with the words. In terms of what some of the general public feel about AI, it's concern, it's caution, it's skepticism. But all that said, if you still have to use AI or still want to get involved in AI, what are the best ways to do it? Nicole Nunguin of the Journal wrote a piece detailing how to get started using AI. She gave away two pieces of advice. One was choose your bot, she said. Think of it as a massive trove of information and learning. She then listed a few. Microsoft's Copilot, anthropic Claude OpenAI's ChatGPT, Google's Gemini. The second one was undo your search brain. Henry. I thought this was good. Instead of searching bikes under $500, which a lot of the times we've put just a couple of words or one word into the search engine and that's how we look for things, provide more detail, she said. So a couple of sentences, as much context as you can. Henry, what's your best advice for folks, especially more skeptical ones, concerned ones, on how to get started with AI?

[Henry Poudley]

That's funny, because when you asked us to prepare on that question, that's exactly what I was thinking. And my suggestion was to change default search engine in your Browser to either ChatGPT or Perplexity for a week. That's something I did and talk about like jumping in the deep end. It was incredibly unsettling because, you know, search and how we search and Google is a verb. Right. We have very deeply ingrained habits around how we search for information. And by just doing that small change, it completely forces you to think differently and the way you ask questions and just everything. And so that would be my piece of advice is just jump in the deep end, change your default engine and explore.

[Marcus]

Yeah, just focusing on that for two seconds. What, I mean, what did, what did you like about experience? What did you not like about it? What were some of the learnings from that?

[Henry Poudley]

Well, sometimes I didn't want a long response. Like, I just wanted the link to the restaurant in town so I can order my takeout. Right. And so, you know, it's, it's interesting. It's not, it's, it's not everything requires A deeply cited, a long summary, but some things, I really found it to be beneficial. I mean, not. Not obviously I'm conflicted as someone who works in publishing, but, you know, having to engage with everything just right on that page and get the information I was looking for was really helpful.

[Marcus]

Yeah, it's kind of similar to how we use different social media platforms for different activities. And it's. It does. See, I was speaking to someone and they were saying, I use Google to find things immediately. I use ChatGPT to get more context and information about a certain thing. And I wonder at what point we'll use whatever service for both and it will know what we're asking and give us the thing we want quickly or with a bit more thought put into it. Grace, how about for you?

[Grace Harmon]

I think going back a little bit, you were talking about that Pew Research about the gap between experts and the general public. And one stat that I really remember from that is just this enormous gap in understanding of how much consumers are actually using chatbots. So I think, you know, if I would have advice for how to get started with AI, a lot of it would be understanding the best use cases. You know, knowing when you should use a chatbot, when you should be using a search engine, you know, if you're coding, what tools are best. And I think that's hard just because a lot of companies, especially OpenAI, have just an enormous number of products in their arsenal and in their lineup. And, you know, it isn't just chat GPT, it isn't just GPT anymore. There's. There's O3, there's O3 mini, there's 4.5. There's just an enormous number of options. So it's difficult, but they are going to have best functions and strengths. So I think, you know, being able to look at what you need to get done, you know, whether it's coding or summarizing or, you know, campaign generation and knowing, you know, where you should be going to and what is going to be best fit for your use. I think that's really, really important.

[Marcus]

That's a good one. In every other part of life, we use the right tool for the right job. And so why would this be any. Any different? Terrific. Well, that's where we have to leave our episode for today, unfortunately. But thank you so much to my guests for hanging out with me today. Thank you to Grace. Thanks, Marcus and Henry.

[Henry Poudley]

Thank you.

[Marcus]

Yes, sir. And thank you to the whole editing crew and everyone for listening into behind the Numbers and Email Marketer video podcast made possible by since. Please subscribe and follow to hear about new episodes and leave a rating and review if you could. We really, really, really appreciate it. Tomorrow you can hang out with Rob Rubin on the Banking and Payment show, where he'll be discussing how banks might be failing customers by not helping them with specific life milestones.