Many of you are here, reading this, because you like the way I write about AI. I’m sceptical, critical, always more interested in what technology is doing to us than what it can do and sometimes, admittedly, I’m a little snarky. But can I tell you a secret? I wasn’t always this way.
For a good few years, I thought AI was mostly harmless, undeniably exciting and probably the future. I wrote about its potential. I went to product launches where AI was the headline feature.
And I still remember one evening about four years ago when I lost several hours to an AI photo app, turning my selfies into a Starfleet officer and then a Blade Runner replicant. Yep, exactly the kind of peak Facebook mum behaviour that makes my eyes roll today.
Just thinking about it makes me cringe. But the reason I’m telling you this is because I really want to normalise something that seems to have become a major taboo. Something that’s always been difficult for humans, but which the internet has turned into a near-cancellable offence: changing your damn mind.
Spend enough time online and you’d be forgiven for thinking the smart, admirable and trustworthy people are the ones who’ve held exactly the same opinion for years. They’ve never backtracked! Or deleted an old tweet! Or admitted they got something wrong!
But the older I get, the more suspicious I become of people who never seem to change their minds at all.
And look, I promise this isn’t another essay about why you should agree with me about AI. I might change my mind again one day about that too.
But the reason I’m using AI as an example here is because it’s become almost the perfect case study for all of this. Few technologies have evolved so dramatically in such a short space of time. Every week seems to bring another breakthrough, controversy or unintended consequence. It’s a field defined by constant change, yet many people still seem anchored to the first opinion they formed.
And that’s what I’m really interested in at the moment. Why do we seem so reluctant to update our beliefs when the world around us keeps changing?
Why changing your mind feels like changing yourself
The easy explanation for why we don’t change our minds is that we tend to ignore evidence we don’t like. Which is largely true, but there’s a whole messy soup of reasons for that.
Looking back, I can see that my early excitement about AI wasn’t really about AI but it reflected how I saw myself.
I’m a technology journalist. I like new things. I like understanding them before everyone else so I can write about them. I enjoy being optimistic about the future. So you can probably see how “AI is exciting” quickly stopped being an idea based on evidence and became part of the kind of person I believed I was.
Psychologists refer to this as identity-protective cognition. Once a belief becomes tangled up with our identity, changing our minds doesn’t feel like just updating our understanding. I guess it’s more akin to dismantling a part of ourselves.
And again, this isn’t unique to AI. You see it in politics, diet culture, parenting and productivity. We adopt beliefs and then build identities around them without even realising it.
The infuriating thing is that simply knowing this doesn’t magically free us from it. I can understand identity-protective cognition and still fall into it. Which brings me to another term I really love: naïve realism.
That’s our tendency to assume we’re seeing reality objectively while people who disagree with us are uninformed, irrational or biased. It’s surprisingly difficult to imagine that we might be looking at the world through different experiences or different evidence. But we’re doing that all. the. time.
Why we cling on anyway
Even when we begin to suspect we’re wrong, changing our minds can come at a massive cost. Whole books have been written about the specifics of these costs, but here are a few that interest me the most.
There’s the simple reality of sunk costs. So, if you’ve spent years defending a position, building a worldview or becoming known as “the AI person” or “the productivity person”, then changing your mind suddenly becomes expensive. Maybe financially, but also socially and emotionally too.
Because again, you’re not simply giving up an opinion here. You’re letting go of the story you’ve been telling yourself about who you are and how the world works.
Psychologist Leon Festinger described something called cognitive dissonance, which I know I mention a lot in these essays, but I genuinely think it’s one of the most useful ideas in psychology.
It’s the tension we experience when our beliefs stop lining up with reality. There are two ways we might resolve that discomfort. So, we can update our beliefs. Or, to make our lives way easier, we reinterpret the evidence so we don’t have to. It’s the psychological equivalent of putting our fingers in our ears and loudly shouting “la la la”.
And a lot of this is about uncertainty. Human beings are so very uncomfortable with not knowing stuff. A familiar belief, even an imperfect one, often feels safer than stepping into the unknown. Letting go of certainty means admitting that, at least for a while, we don’t have all the answers. And that’s not a particularly comfortable place to be.
But our beliefs don’t just exist inside our own heads either. For most of human history, belonging to a group wasn’t optional. It was necessary for survival. So agreeing with the group often brought acceptance and status and security. But disagreeing could mean friction, rejection or even exclusion.
That’s one reason why changing our minds can feel surprisingly threatening sometimes. We might be constantly weighing the potential social consequences of accepting something new.
Some researchers suggest that when our deeply held beliefs are challenged, the brain can respond in ways that resemble responses to physical threats. Whether or not the comparison is exact, the feeling is familiar. We become defensive, stop listening and look for evidence that confirms what we already believe. So even curiosity becomes so much harder when our nervous system shifts into self-protection mode.
And, completely unsurprisingly at this point, the internet has amplified every one of these instincts. We know that algorithms reward certainty, confidence and consistency far more than curiosity or nuance. And someone who publicly changes their mind risks looking weak or hypocritical.
In his excellent book, How Minds Change, David McRaney tells the story of Charles Veitch. Veitch became well known for promoting 9/11 conspiracy theories before later really dramatically changing his mind after examining more evidence and talking with some of the people directly affected by the tragedy.
McRaney uses Veitch’s story as a really solid case study in how the most extreme views can change. But unfortunately, it’s also a really devastating case study in how the people around you can become extremely hostile when you change your mind. Rather than his new views being welcomed, he was met with abuse, threats and rejection from the very community that had once embraced him.
Why AI amplifies all of this
Everything I’ve written so far applies far beyond AI. To parenting, politics, relationships, health and fitness. We cling on to all sorts of beliefs far longer than we should.
But look, I keep coming back to AI because it seems to magnify almost every one of these psychological traps at once.
Technology moves incredibly quickly, yet our identities don’t. Many of us formed an opinion about AI years ago, when ChatGPT was writing awkward emails and turning us into Pixar characters. Since then we’ve seen convincing deepfakes, emotional dependency, sextortion scams, questions over creative work, environmental concerns, surveillance through smart glasses and entirely new ways of interacting with machines.
Now, whether those developments make you more optimistic or more sceptical isn’t really the point. It’s that the evidence keeps changing.
If the evidence changes but our opinions don’t, I think it’s worth asking why. I don’t think the answer is that one side is rational and the other is irrational. I think we’re all vulnerable to becoming attached to the first story that made sense to us.
I’ve noticed something else too. In technology especially, excitement can become a sort of moral virtue. Questioning a new technology isn’t always treated as curiosity. Sometimes it’s actually ridiculed as fear or pessimism or evidence you’ve become a Luddite or you’ve somehow fallen behind.
I recognise that instinct because I’ve felt it myself. For years, part of my job was finding exciting new technologies before they became mainstream. So there was always this subtle pressure to stay optimistic. To be the person who spotted the future first. At some point, I realised I’d absorbed the idea that questioning technology meant becoming anti-progress in some way, right?
I don’t believe that anymore. In fact, I think the opposite is true. Progress, to me at least, depends on asking difficult questions. If we decide that criticism automatically makes someone anti-technology, we start judging the people asking the questions. And that’s exactly how many big tech companies get away with so much unchecked. That’s a dangerous place to end up. And I hate to say it, but I see a lot of other tech journalists there right now.
While researching all of this, I came across a phrase I wish we’d hear far more often: epistemic humility. It’s the recognition that our knowledge is always incomplete and we might be wrong. It’s what I’m aiming for.
And look, I can see how this could be criticised. But it doesn’t mean having no convictions or changing your mind every Tuesday afternoon because someone disagreed with you on LinkedIn. It means holding your convictions with a little more detachment so that new evidence can still change them.
Why this matters far beyond AI
My own opinion about AI didn’t change overnight. It changed slowly, one interview, story and deeply uncomfortable experience at a time.
There were a couple of key moments that stand out in being pretty pivotal. Like, being the victim of AI image abuse. Reporting on AI image abuse. Speaking to people who’d become emotionally dependent on chatbots. Learning more about the environmental costs. Watching story time after story time of people being filmed without their consent with AI glasses. And watching heart-breaking interview after heart-breaking interview of the people who have to live next to data centres. There’s so much more, but they’re the main ones.
And I think the reason I’ve become so fascinated by this recently, why I’m sat here writing this essay when I really should be packing for a holiday, is that the willingness and ability to change your mind doesn’t only matter when we’re thinking about technology. I believe it matters in everything. Our relationships, work, politics. In the stories we tell ourselves about who we are and who other people are.
I can say confidently now, after feeling like I’ve lived several distinct lifetimes within my one lifetime so far, that every part of life asks us, sooner or later, to let go of ideas that no longer fit reality.
I don’t think we really benefit from having stronger opinions or becoming perfectly informed. I think we benefit from becoming more comfortable with uncertainty, less defensive when we’re wrong and far more willing to notice when reality has moved on without us. Because every day new evidence appears, things change, technology evolves, we meet different people and, maybe most importantly, we become different people too.
My morning listen whilst I go about more mundane things