1.
There’s a psychological concept that’s dark, confronting but incredibly fascinating: moral disengagement. It asks how seemingly good people can do bad things.
How can “normal” people commit violent crimes? Why do people follow orders and not think about the outcomes? In more everyday ways, how do we convince ourselves that something isn’t really our responsibility? That everyone else is doing it? That we didn’t have a choice? And that our own actions are so small they wouldn’t make a big difference anyway?
I told you, this is confronting.
But many would argue it’s imperative we better understand why this happens. In order to prevent it and maybe also remain vigilant to the tiny ways it can show up so we can better spot it, in the people we choose to have in our lives, and even ourselves.
The reason I’m writing about this today isn’t because I think our tech use or our AI opinions can be compared to committing violent crimes in any way. But moral disengagement doesn’t always show up in extreme situations. The same sorts of rationalisations are present in some everyday decisions, especially when our behaviour starts to conflict with what we believe about ourselves.
2.
The big AI news over the past few weeks has been that companies are introducing new ways to detect AI writing and watermark AI content.
Now, there’s a lot of talk about whether this is good or bad. Whether detection is really accurate. If people will be falsely accused of using AI. Whether some of these measures aren’t really about detection at all, but actually serve the interests of AI companies by feeding training data back in.
But as always here, I’m more interested in the human side of this tech news. And over the past few weeks, I’ve been watching a lot of people crash out about the fact AI detection is being rolled out.
Now, there have been a lot of different reactions, they’re not all the same. Some are concerned it’s not accurate. Others are worried they’ll be falsely accused of using AI. Others vaguely talk about how bad witch hunts are. And all of these opinions are valid.
But given how many people we know use AI, there must also be people among those reacting who are worried for a much more personal reason. Which is that they use AI to create things and they’re worried they’ll be found out.
3.
And you know what? I empathise. Can we really wholly blame the user for using AI? We have been led to believe these tools are time-savers and life improvers.
And, as a result, some people have leaned on them for everything from writing their LinkedIn posts to life advice to strategic thinking at work to connection and everything in between. Among those, I suspect, many people are incredibly burned out and simply wanted to claw back some time.
(And if you’re new here and think I’m starting to sound a bit judgey, know that I went through a stage of using AI a little too.)
What’s more, some people are pressured to use it at work. I particularly feel for others who may desperately want to prise themselves away from chatbots but feel dependent on them now to write, to sort through advice, to feel connected.
Is it really any different from blaming people for being addicted to social media when those platforms have been so carefully engineered to keep us coming back?
I think this provokes really interesting questions about how much we can say: “I did this because I was manipulated. It was forced on me. I was tired and burned out. My boss, in a fit of what might be AI psychosis, forced me to use it.” Rallying against that when you need a job is hard.
But does that let people off the hook? I personally think it’s normal and human to sometimes fall for hype. And maybe it’s not your fault you did, but it is your responsibility to keep thinking critically and assessing whether you’re playing an active part in something that isn’t good for you, your brain, humanity and the planet.
4.
So let’s talk about the theory of moral disengagement from psychologist Albert Bandura.
Bandura was interested in how regular people can do terrible things without feeling like terrible people. He wanted to know how people explain away behaviour that doesn’t align with their values, how we convince ourselves what we’re doing is fine and how we shift responsibility elsewhere.
Again, to make it clear, I’m not saying AI is comparable to committing violent acts or going to war, obviously it isn’t. But I do think the psychological mechanisms that Bandura identified can give us an interesting framework for thinking about responsibility in the more ordinary parts of our lives.
Because it turns out we’re really good at explaining away behaviour that doesn’t align with our values when it suits us.
In this context, that’s stuff like: “everyone is doing it, my boss expects it, AI is inevitable, my individual use makes no difference.” Or, the one that irks me the most: “it’s just a tool.”
Bandura talked about the displacement and also the diffusion of responsibility. That it can even be pushed on someone else or spread across a group until the part we play starts to feel really small.
I can see this happening with AI. I’ve seen people justify their use by saying, “if everyone is using it, does my use really matter? If my boss expects it, is it really my decision? If these companies are building it into everything anyway, what difference does it make whether I use it or not?”
Another one I’ve seen a lot at the moment around AI glasses is, “if there are lots of cameras around anyway, why does it matter that I use one on my face to record you? Millions of people already have Meta glasses, right?”
And, as I said before, some of those explanations might be legitimate. I’ve heard a lot of first-hand stories about people being pressured to use AI at work. Someone might be burned out. Someone might work somewhere that’s decided AI is the future and refusing to use it could put their job at risk.
But someone can be genuinely pressured to use AI at work and then that pressure can then become a reason to avoid examining our own use of it.
At what point does explaining why we do something become a way of absolving ourselves of any responsibility for doing it?
5.
Which brings me back to AI detection. Because there’s something a bit strange you might have noticed about this whole debate if you’ve also been following it online.
People who write and generate things with AI would, in many cases, be fine if they admitted it.
Substack now has AI detection, as just one example, and if it says you used AI, you can write a little statement about how and why you used it. Maybe you just used it for the intro. Maybe you got it to proofread something. Maybe you dictated all of your own ideas and it just wrote the whole thing.
The interesting part here for me is when someone does use AI to create something, presents that work as entirely their own and then doesn’t want people to know.
Let’s get back to that idea of diffusion of responsibility because it’s essential here. That’s when responsibility is spread across a group. When that happens it becomes way easier to minimise our part in what’s happening.
And anonymity really helps with that. Bandura says that group action can make it easier to diffuse responsibility because you can always see other people as being more responsible than you are or think they’re all doing it so what difference do your individual actions make.
That feels so very relevant here: “Everyone uses AI and no one will know I have, so what’s the harm?” That’s way easier to stomach than: “I used AI to make this.”
Detection forces us to reckon with that. It takes something vague, collective and, most importantly, very easy to minimise and makes it individual. What did you use? What did you make? What did you allow other people to believe that you made?
What I also find really interesting is that Bandura talks about how denying or minimising the consequences of our behaviour is what might prevent our usual moral self-regulation from kicking in.
And I wonder whether that’s what makes the prospect of detection so uncomfortable. Some of those ways of minimising your own responsibility become harder when your individual behaviour is suddenly in the spotlight.
6.
I have a lot of empathy for how it might feel to have been using AI and not questioning it all that much and now suddenly to feel like you’re facing a sort of reckoning. But I’m not gleeful about the fact people are probably crashing out for very confronting and legitimate reasons.
I blame the tech companies more. The hype mongers, the celebs with endorsements, the influencers, the people who spent years making this tech seem irresistible and inevitable.
But I also don’t think I can write all of this about diffusion of responsibility and then put all of the responsibility there.
We might be under enormous pressure and influence. But there are still moments when we can stop and ask what we’re participating in.
And you know what? The thing is, as I’ve said before many times here, I might be wrong about all of this. Maybe AI use will turn out, despite all of the issues right now, to be a good thing, a net positive thing.
But even if that’s true, wouldn’t it be better to spend some time now assessing whether you really do want to use it? Whether you’re comfortable with what you’re using it for? Whether you’re taking responsibility for that choice rather than assuming everyone else is doing it and it’s inevitable?
Whether you might be participating in something you’ll later wish you’d spent way more time thinking critically about.
Because see, as I get older, and I hope wiser, I realise that I am not sure about anything because nothing is ever really certain. So, all I can keep doing is showing up again and again, questioning, wondering and taking responsibility when I can. Even if it’s confronting, especially if it’s confronting.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.