When morality meets the machine

Summary

---

Podcast Summary: Click Here – "When morality meets the machine"

Host: Dena Temple Raston
Guest: Shannon Valor, Professor of Ethics and Artificial Intelligence, University of Edinburgh
Date: March 6, 2026

---

Episode Overview

This episode of Click Here delves into the profound question: What happens when we outsource our toughest moral decisions to artificial intelligence? Host Dena Temple Raston and ethicist Shannon Valor explore how AI’s increasing role in everyday decision-making—from relationships to justice—may erode our moral agency and reshape humanity’s understanding of right and wrong. The discussion spans the risks of trust in "superhuman" machines, the dangers of offloading moral judgment, and the critical need to preserve human moral skills in the digital age.

---

Key Discussion Points & Insights

1. The Allure of Outsourcing Morality to Machines

• The episode opens with the fundamental question: What happens to us if we let AI step in when decisions are hardest and stakes are highest? (00:02)
• AI’s role expands: Increasingly, AI assists with issues humans traditionally grappled with—big questions about relationships, jobs, and even spiritual meaning.

2. Exaggerating AI’s Capabilities & Eroding Human Agency

• Valor: "A lot of the rhetoric around AI… is to use this framing as a way to justify diminishing human agency, sidelining human creativity, power, judgment, and responsibility." (03:27)
  • Not only is AI’s ability oversold, but this narrative quietly encourages society to defer to machines, undermining our own capacities.

3. The Danger of “Superhuman” Language

• Valor on performance: Speed or efficiency does not equate to wisdom or ethical reasoning.
  • "The language very quickly slipped into describing these as superhuman… and I found this language really dangerous from the beginning and misleading." (04:12)
• Temple Raston: This shift in language reframes AI as something inherently superior, making us more likely to trust it by default. (04:37)

4. What AI Actually Does: Reflection Not Judgment

• Valor: "Ethics is not a math problem." (05:34)
  • AI can only optimize for predefined, mathematical targets; it cannot fundamentally understand or weigh ethical dilemmas.
• Temple Raston: Large language models merely predict the most likely sequence of words, acting as "less like a compass and more like a mirror." (06:06)
• Valor’s mirror analogy: AI reflects humanity’s combined digital output—not guidance, but synthesis of what’s been fed into it. (06:46)

5. Seeking Wisdom or Shortcut?

• People increasingly seek out AI for help—not because it’s wise, but available and willing to offer answers. (07:26)
• Temple Raston: "What it gives back isn’t really a way forward. It's more like a summary of where humanity has already been." (07:26)

6. The Loss of Vision & Moral Creativity

• Valor: "A system trained on the past can’t imagine a future that breaks from it." (09:47)
  • AI lacks "moral vision" and creativity fundamental to social progress.
  • If AI made policy in the 19th or 20th century, "I wouldn’t have the right to vote. I’d still be considered property." (09:06)
• Temple Raston: History shows progress comes from messy conflict, risk, and redefining boundaries—things machines cannot replicate. (08:25)

7. Sycophancy and Emotional Safety

• Valor: AI chatbots are "fine-tuned to please the user because that's what's commercially viable…to keep the user happy." (10:44)
  • Example: AI giving relationship advice may encourage people to always see themselves as victims ("sycophancy"), rather than prompting honest self-examination. (10:09)
• Temple Raston: Real human advice involves risk and vulnerability; with machines, there is none. (11:01)

8. Atrophying Moral Muscles

• Valor: Evidence shows even experienced professionals can see “skills degrade…if they’re offloading much of that judgment to AI systems.” (16:34)
  • If we stop practicing moral decision-making, those skills weaken, just like muscles. (17:00)
  • This loss ripples outwards, threatening the very foundation of society. (17:26)
  • "It's fundamental to the relationships…that we all be able to open ourselves up to other humans." (17:35)

9. Where AI Can—and Should—Help

• Temple Raston: There are real advantages to using AI for scale, e.g., "tracking supplies, beds, and who's where" in an emergency room, when used to support—not replace—human judgment. (18:08)
• Valor: "I think there’s a world where the machines are in fact helping us make these kinds of decisions in scenarios where we simply can’t make the best decisions on our own." (18:32)
• But accountability remains an intractable problem. Who answers for a machine’s life-altering errors? (18:43, 19:01)

10. The Lessons of Social Media: Drift, Regret, and Degraded Institutions

• Valor: "Relatively few of us would think that our political environment is better for it or that our media environment is better…we’ve settled for a worse condition…in part from not choosing to regulate these technologies appropriately." (20:02)
• Temple Raston: "A slow drift followed by regret." (20:29)
• If we repeat the same systemic inaction with AI, the damage to our moral identities could be enduring. (20:37)

11. Recovering Morality: Urgency and Hope

• Valor: "We’re at a very critical moment in time…a window that has not closed to reclaim our moral agency…our ability to work with technology differently. I’m always…hopeful, in the long run, because humans have a long history of wiggling ourselves out of difficult predicaments." (20:54)
• But hope alone is not enough—meaningful action and conscious design choices are required.
  • "The kind of AI we have and what we let AI become in our lives and in our society is still very much up to us." (21:35)

---

Notable Quotes & Memorable Moments

• "What happens when we start outsourcing the most important part of being human?"
  — Dena Temple Raston (00:02)

• "We can forget about all of those aspects of ourselves."
  — Shannon Valor on how AI rhetoric diminishes human agency (03:27)

• "Ethics is not a math problem."
  — Shannon Valor (05:34)

• "A system trained on the past can’t imagine a future that breaks from it."
  — Dena Temple Raston (09:47)

• "If AI had been doing the policy making…in the 19th or 20th century, I wouldn’t have the right to vote. I’d still be considered property."
  — Shannon Valor (09:06)

• "They’re fine-tuned to please the user because that’s what’s commercially viable."
  — Shannon Valor on AI chatbots (10:44)

• "So we've got some evidence that even experienced skilled professionals can see those skills degrade…if they're offloading much of that judgment to AI systems."
  — Shannon Valor (16:34)

• "I think we're at a very critical moment in time where we have a window that has not closed to reclaim our moral agency…"
  — Shannon Valor (20:54)

• "Machines can give answers, but they can't replace the work of being human."
  — Dena Temple Raston (21:44)

---

Key Timestamps

| Time (MM:SS) | Segment / Quote | |--------------|------------------------------| | 00:02-00:36 | Dina Temple Raston frames the episode’s core questions on AI and morality | | 03:27 | Valor discusses AI rhetoric diminishing human agency | | 05:34 | Valor: “Ethics is not a math problem.” | | 06:46 | Valor’s “mirror” analogy for LLMs | | 09:06 | Valor: On historical moral progress and AI’s limitations | | 10:44 | Valor: AI algorithms prioritize user satisfaction (sycophancy) | | 16:34 | Atrophy of moral skills when moral decisions are outsourced | | 17:35 | The societal consequences of lost moral capabilities | | 18:43 | The accountability problem in AI-driven decisions | | 20:02 | Parallels to social media’s slow regretful drift | | 20:54 | Valor: The critical moment for reclaiming moral agency | | 21:35 | Valor: Our future with AI remains our choice |

---

Conclusion

The episode offers a nuanced, urgent exploration of what’s at stake as people hand over increasingly complex, morally laden decisions to algorithms. While recognizing the legitimate utility of AI in certain contexts, Dena Temple Raston and Shannon Valor warn that a future where humans are sidelined in matters of ethics is neither inevitable nor desirable. Instead, they call for intentional choices—preserving space for human judgment, risk, and moral creativity—so that technology supports, rather than replaces, what makes us truly human.

---

Transcript

[Dena Temple Raston]

From recorded future news and prx, this is click here. What happens when we start outsourcing? The most important part of being human? If we let AI step in when decisions get hardest, when the stakes are highest, and what does that do to us?

[Shannon Valor]

I think it's going to be a while before we actually know what effect they've had on our moral skills and capabilities.

[Dena Temple Raston]

I'm Dena Temple Raston and this is Click Here. We tell true stories about the people making and breaking our digital world. Today we continue to explore what happens when we hand our moral judgment over to machines. To help expand the conversation, I talked with Shannon Valor, professor of Ethics and Artificial Intelligence and at the University of Edinburgh and a researcher for the Edinburgh Futures Institute. Among the things we discussed is if we turn to AI to help us make hard decisions, what happens to us when we let it, and what happens to our sense of right and wrong?

[Shannon Valor]

By the time we have that evidence, the effects will have already transpired at scale.

[Dena Temple Raston]

Stay with us.

[Sponsor/Announcer]

This show is supported by Odoo. When you buy business software from lots of vendors, the costs add up and it gets complicated and confusing. Odoo solves this. It's a single company that sells a suite of enterprise apps that handles everything from accounting to inventory to sales. Odoo is all connected on a single platform in a simple and affordable way. You can save money without missing out on the features you need. Check out [e-mail] that's O D O

[Dena Temple Raston]

O.com support for click here comes from Servil. IT teams waste so much time on repetitive tickets. All those password resets, access requests and onboarding. With Servil, you can cut 80% of that busy work. So all it has to do is write what they need in plain English and Servl makes it happen instantly. Consider onboarding new hires waiting around for days, managers asking for approvals. It gets pulled away from meaningful work. With Servl, a manager can simply request onboarding with a quick slack message and just like that, access happens in seconds automatically with all the right approvals. It never even has to touch it. If I were starting a tech company, servl would be a must have. It saves time and money and lets it focus on actual problems. That's why Servol powers the fastest growing companies in the world like Perplexity, Mercore, Furkada and Clay. Get your team out of the help desk and back to the work they enjoy. Book your free [e-mail] clickhere that's S-E-R-V-A-L.com clickhere. When a new tool starts showing up in places where humans had previously been weighing right from wrong, it's worth asking, what's it changing in us?

[Shannon Valor]

A lot of the rhetoric around AI is not just about exaggerating the power and capabilities of AI, But I think the deeper and more dangerous trend here is to use this framing as a way to justify diminishing human agency, sidelining human creativity, power, judgment, and responsibility. Essentially, what we're saying is we can forget about all of those aspects of ourselves.

[Dena Temple Raston]

Shannon has spent years thinking about how technology quietly reshapes what it means to be human. And one thing that worries ethicists is how easily we start to talk about machines as if they're more than tools.

[Shannon Valor]

We can see that the machine can perform it 100 times faster than any human on earth could. But what happened is that the language very quickly slipped into describing these as superhuman, divorced from a measurement of a particular task at a particular target, like speed. And I found this language really dangerous from the beginning and really misleading, because

[Dena Temple Raston]

once we start talking about machines as superhuman, something subtle shifts. It no longer sounds like something we built. It sounds like something better than us, something we might trust even when we shouldn't. And you can hear that shift everywhere now, in headlines and social feeds and the kinds of problems people are starting to hand over to AI.

[Shannon Valor]

Using AI chatbots as personal therapists is on the rise.

[Dena Temple Raston]

Your kids might be having romantic relationships with artificial intelligence.

[Mike Pesca]

Here's a look at what Chad GPT had to say after a user asked for some. Some help with the female that he was interested in.

[Dena Temple Raston]

ChatGPT AI spits out answers in language that sounds human, empathetic, thoughtful, reassuring. So it's easy to feel like it understands, but it doesn't. AI doesn't reflect, it doesn't care. It doesn't weigh right and wrong. It just calculates.

[Shannon Valor]

The way you design machine learning models to optimize for what we call an objective function is you essentially set up the goal as a math problem, and you say, this is how we want you to calculate success, and then we want you to learn how to return results that come as close as possible to this target, which again, has to be mathematically predefined in order for the machine to know what a good answer looks like. And ethics is not a math problem.

[Dena Temple Raston]

So, in plain English, large language models don't think about your question. They predict what words are most likely to come next by drawing on what humans have already written. The Good, the bad, and the deeply online. From academic papers to bad Reddit takes to fanfiction. So instead of looking forward, it's actually looking sideways across the past and handing that back to you less like a compass and more like a mirror. Can you talk about what AI is actually reflecting back to us?

[Shannon Valor]

If you think about the way mirrors work, they reflect wavelengths of light that are falling on glass and then coming back to us, and therefore we get the reflected image. There is a very similar thing that's happening with large language models, which is that they are ingesting large volumes of data. But instead of light falling on glass, you have huge troves of data falling on algorithmic surfaces. These are sort of mathematical instructions for how to respond to data and how to analyze data.

[Dena Temple Raston]

People are already turning to AI for help with life's biggest questions. Who should I date? Should I change jobs? And what does this mean spiritually? Not because it's wise, but because it's available and it answers. The problem is, and we mentioned this before, what it gives back isn't really a way forward. It's more like a summary of where humanity has already been. The greatest hits, the worst takes. It's a record, not a roadmap.

[Shannon Valor]

Think about climbing a difficult mountain you that you've never, never summited before, and you have no idea what the terrain looks like. You would not navigate that mountain with a mirror pointing backwards. That just showed you where you'd already been and the routes you'd already taken. The ability to navigate uncertain and new futures is fundamental to, to human development.

[Dena Temple Raston]

And humans have done this before, moved forward without a clear map. They do it mostly by stumbling and arguing and protesting, getting it wrong and then trying again. That messy friction is how societies change. It helped end slavery and segregation. It put limits on nuclear weapons. It won women the right to vote. None of that was efficient. None of that was optimized. Instead, it required courage, imagination, and a willingness to step into the unknown without guarantees. Shannon worries about what happens if, instead of struggling through moments like those, we start looking to a machine for the answer.

[Shannon Valor]

Instead, if AI had been doing the policy making and doing the voting and making the moral decisions in the 19th century or the 20th century, I wouldn't have the right to vote. I'd still be considered property. That was a leap forward into an unknown political and moral space.

[Dena Temple Raston]

It's vision, right?

[Shannon Valor]

Yes. It's moral vision. It's moral creativity. And this is a fundamental gap in these systems capability,

[Dena Temple Raston]

A system trained on the past. Can't imagine a Future that breaks from it. It can reinforce patterns we've already decided were unjust. And it can't absorb the full weight of human experience, the history, the content, the fear or hope that never gets typed into a prompt. And it isn't built to challenge us.

[Shannon Valor]

People are getting advice about how to handle marital arguments. And there's been some reporting on this, and it's quite disturbing, right, the way that some people have found that the machines are encouraging actually them to see themselves as the victim in a relationship where they may be as flawed as their partner. And so these technologies have what we call sycophancy.

[Dena Temple Raston]

Sycophancy, a technical term for something very human. Feeling, flattery.

[Shannon Valor]

They're fine tuned to please the user because that's what's commercially viable, right? The machine has to keep the user happy or else the user stops using it and stops paying for it.

[Dena Temple Raston]

Healthy relationships aren't built on constant agreement. They involve honesty and friction and repair. And when we turn to other humans, friends, pastors, partners, something else enters the equation. Risk vulnerability.

[Shannon Valor]

Traditionally, one of the important pastoral functions of a religious leader was to listen to their parishioners, troubles without judgment, with compassion and in a way that would provide guidance rather than condemnation. But humans are flawed and we're imperfect and we do judge. So I think people are taking a risk when they talk to another human. You're being vulnerable. You're opening yourself up to another human being whose professional vocation may require that they treat you with compassion and sympathy and not judgment. But you'll never know whether that's in fact what you're getting. Right.

[Dena Temple Raston]

That uncertainty, the emotional risk, is part of what shapes moral judgment. And a chatbot removes that risk entirely.

[Shannon Valor]

If you go to a machine, you know that you're not being judged because the machine is incapable of making that kind of deep, personal moral judgment of your character. It can pretend to do it, but we know that the machines are all dark inside. There's no thought, there's no perception, there's no feeling towards you.

[Dena Temple Raston]

And that's part of the appeal. No judgment, no discomfort. But what happens when we start preferring that? That's next after the break. Support for Click Here comes from Monarch, the all in one personal finance tool designed to make your life easier. Lots of us are thinking about our finances in 2026. But when it comes to paying off debt, building an emergency fund, and saving for major milestones, you need a tool that doesn't just track your wallet. You need something that helps you plan, project, and achieve your goals. Set yourself up for financial success this year with Monarch. It brings your entire financial life, budgeting, accounting and investments, net worth and future planning all together in one dashboard. It's cleaner than a big spreadsheet documenting all your expenses, which can make you feel bad about past spending. Monarch keeps you focused on planning ahead and gives you the complete picture so you can make decisions that actually move the needle. Set yourself up for financial success in 2026 with Monarch, the all in one tool that makes proactive money management simple all year long. Use the code [e-mail] for half off your first year. That's 50% off your first [e-mail] with the code clickhere Support for Click here comes from Quince these days I'm all about quality over quantity, especially in my closet. If it's not well made and versatile, it's just not worth it to me. That's why I love Quince. The fabrics feel elevated, the cuts are thoughtful, and the pricing actually makes sense. Quince makes high quality wardrobe staples using premium fabrics like 100% European linen, 100% silk and organic cotton poplin, and they come directly from safe, ethical factories. They cut out the middleman so you don't pay extra for brand markups. It's just quality clothing at a good price and it's consistently rated 4.5 to 5 stars by thousands of customers. My new favorite sweater? My Quince Cashmere quarter zip. I actually find excuses to wear it. It looks great, super soft and it's one of those classic pieces you keep going back to. Right now, if you go to quince.com clickhere you can get free shipping and 365 day returns. That's a full year to wear it and love it. And you will. Now available in Canada. Don't keep settling. For clothes that don't last, go to Q-U-I-N-C-E.com clickhere for free shipping and 365 day returns. Quince.com clickhere hi, I'm Mike Pesca, host

[Mike Pesca]

of the Gist, the longest running podcast on news and analysis in the business. Every day I do, I think, a fascinating interview, but I also do a spiel where I will challenge you because I hate listening to podcasts that just agree with each other and only agree with me. So I'm going to say 25% of the time you'll say, mike, I think you're wrong and 75% of the time you'll say, great point. Well said. Now those numbers might be reversed for some people, they probably shouldn't listen. Check out the gist every single day for a decade now. It is called Internet.

[Dena Temple Raston]

I use the World Wide Web.

[Shannon Valor]

Information superhighway.

[Dena Temple Raston]

Cybersecurity. Why do things go viral?

[Shannon Valor]

Click here.

[Dena Temple Raston]

Part of being human is that we often try to find the path of least resistance. The shortcut or quick hack that gets us to the solution we need without having to spend all that energy navigating some pesky, complicated problems. But what happens to those complex skills if we don't use them? Shannon says they weaken.

[Shannon Valor]

So we've got some evidence that even experienced skilled professionals can see those skills degrade in a relatively short period of time if they're offloading much of that judgment to AI systems.

[Dena Temple Raston]

So what happens if we stop making those calls ourselves? Not just the big moral crossroads, but the small decisions that quietly shape who we are?

[Shannon Valor]

What happens to our capacity, our ability to retain moral wisdom, Our ability to retain the ability to make tough moral calls? Because if moral capability is developed through habit and practice, then it can be lost or degraded through neglect, just in the way that our other cognitive faculties can.

[Dena Temple Raston]

If that capacity weakens, the effects don't stay personal. They ripple outward into families and institutions and societies.

[Shannon Valor]

Frankly, the foundations of society fall apart at that point. And it's very easy for it to feel too scary if your whole life you've had a machine that promises to give you the same kind of feedback without the risk. It's important that we take risks and make ourselves vulnerable to one another. It's fundamental to the relationships of family and friendship and community that we all be able to open ourselves up to other humans.

[Dena Temple Raston]

Shannon isn't arguing that AI has no role at all. There are areas where machines can genuinely help, especially when the challenge is scale and not judgment. In an emergency room, for example, tracking supplies, beds and who's where, tools that support human judgment rather than replace it.

[Shannon Valor]

I think there's a world where the machines are in fact helping us make these kinds of decisions in scenarios where we simply can't make the best decisions on our own.

[Dena Temple Raston]

But even there, another problem emerges. Accountability. If a machine makes a life altering decision about your health or your freedom or your future, who answers for it? Who can change it? You can't vote an algorithm out of office and you can't cross examine one in court.

[Shannon Valor]

And yet I hear every day people talk about how wonderful it's going to be when human policymakers and human judges are letting their decisions be guided by super intelligent superhuman AIs. And that is A future that we should resist with every fiber of our being.

[Dena Temple Raston]

And this isn't hypothetical. Millions of people now use AI every day. If even a small fraction of them experience harm without a clear way to repair it, it won't stay isolated. And what's more, we may not know right away that it even happened.

[Shannon Valor]

I think it's going to be a while before we actually know what effect they've had on our moral skills and capabilities. And that's kind of scary, because by the time we have that evidence, the effects will have already transpired at scale.

[Dena Temple Raston]

We've seen this movie before with social media.

[Shannon Valor]

We have settled into a pattern. But relatively few of us would think that our political environment is better for it or that our media environment is better for it. What instead we've done is settle for a worse condition of those institutions as a result, in part of not choosing to regulate these technologies appropriately.

[Dena Temple Raston]

A slow drift followed by regret.

[Shannon Valor]

If we follow the same pattern with AI, then I think we will end up settling for a degraded version of ourselves and our societies.

[Dena Temple Raston]

Because once those moral muscles atrophy, it's not clear how long it takes to rebuild them.

[Shannon Valor]

I think we're at a very critical moment in time where we have a window that has not closed to reclaim our moral agency, our moral responsibility, and our moral intelligence and our ability to learn to work with technology differently. So I'm always, shall we say, hopefully, in the long run, because humans have a long history of wiggling ourselves out of difficult predicaments that we've put ourselves in.

[Dena Temple Raston]

But hope, she says, isn't automatic. It requires action.

[Shannon Valor]

Now, the kind of AI we have and what we let AI become in our lives and in our society is still very much up to us.

[Dena Temple Raston]

Machines can give answers, but they can't replace the work of being human. And in a real crisis, there may be no time to ask an AI what to do. There'll only be you, your instincts, your values, your judgment. So at the end of the day, it's worth asking, who's really more superhuman, the machines or us? This is Click. Click. Here is a production of Recorded Future News and prx. Today's show was written and produced by Megan Dietre, Sean Powers, Erica Gaeda, Zach Hirsch and Casey Georgie. It was edited by Karen Duffin and Sarah Covedo and Fact Checked by Darren Ancrum. Original music is by Ben Levingston with additional music from Blue Dot Sessions. Our staff writer is Lucas Riley, our illustrator is Megan Gough, and our sound designers and engineers are Jake Cook and Jesse Niswonger. Find us on X or Facebook @ClickHereShow or leave us a voice message at @6615CH. Talk. Sometimes we'll turn those moments into reporting, sometimes into a conversation, and sometimes into a future story you'll hear on this show. I'm Dena Temple Raston, and thanks for listening.

[Sponsor/Announcer]

Support for this program comes from Recorded Future. In cybersecurity, the biggest risk isn't what can be seen, it's what gets missed. Recorded Future analyzes billions of signals to help organizations stay ahead of threats. Recorded Future Know what matters? Act first.

[Recorded Future Representative]

Looking for more of the cybersecurity and intelligence coverage you get on? Click here, then check out our sister publication, the Record from from Recorded Future News. You'll get breaking cyber news from reporters in New York, Washington, London, and Kyiv, among others, and you'll see for yourself why it attracts hundreds of thousands of page views every month. Just go to the Record Media.