When I ask a room whether AI will disrupt their industry, nearly every hand goes up. But if I ask that same room whether AI will disrupt their own specific job, the hands mostly stay down. I've watched this happen enough times that it stopped feeling like a coincidence. There's a name for it. Researchers call it invulnerability bias, a fancy way of saying you can believe a risk is real and still assume it somehow applies to everyone except you.

Why Invulnerability Bias Feels So Reasonable

Part of what makes invulnerability bias so easy to miss is that most people have good reasons for believing their own job is different. You know how much experience your work requires, which relationships you rely on, and what you do that never appears in a job description. Then you hear about AI disrupting another industry or changing someone else's role, and it can seem perfectly reasonable to assume your situation is different.

A study published last year in Scientific Reports put real numbers behind this. On average, people rated their own jobs as less exposed to AI than jobs in general, although the strength of that bias varied by profession. Pew found a similar gap at a national scale. Sixty-two percent of U.S. adults think AI will have a major impact on workers generally, while only 28% think it will have a major impact on them personally. That's a lot of people looking at the same disruption and assuming it's happening to somebody else's career.

Who Actually Carries Invulnerability Bias

What I found most interesting was who was least likely to have the bias. People who reported knowing more about AI tended to show less invulnerability bias. The people who were more familiar with the technology were also more willing to see how it could affect their own work.

Profession made a difference too. Bias ran highest in healthcare, law, and public administration, and lowest in technology, engineering, and architecture, where only about a third of people showed it. I understand why someone practicing law or medicine might feel confident about the future of their work. They spend their days exercising judgment and drawing on specialized expertise that has historically been difficult to automate. That confidence can feel earned rather than assumed.

What I find most useful is the connection between AI knowledge and the bias. The more people reported knowing about AI, the less invulnerability bias they tended to show. That makes me wonder how much confidence about your own job comes from actually understanding the technology and how much comes from what you've heard about it secondhand.

How Invulnerability Bias Plays Out At Work

Forrester found the same pattern inside B2B marketing. Eighty-two percent of respondents agreed AI would automate marketing work people currently do, while only 15% thought it would largely replace their own job. An entire industry can see the disruption coming while nearly every individual inside it assumes their own job is somehow further from it.

That contradiction rarely gets flagged inside a company, since a leadership team can commission an AI strategy, present it to the board, and fund it, while most of the people actually executing that strategy privately believe it describes someone else's role.

Part of why this is so easy to miss is that most jobs don't change all at once. AI rarely changes your entire role starting Monday morning. It begins taking over pieces of the work. You might notice a report gets drafted faster or research that once took hours takes minutes. Each change looks small enough on its own to shrug off, so it's easy to conclude your job has basically stayed the same. Over time, those small changes can add up to something much bigger.

AI also tends to get discussed in terms of millions of jobs and entire industries rather than individual jobs. That makes the change feel enormous in general and strangely distant up close. You can believe the statistics completely and still feel, sitting in your specific chair doing your specific work, like the exception they never quite meant to include.

Where Curiosity Fits Into Invulnerability Bias

I think about this as a curiosity problem as much as a confidence problem. Genuine curiosity means asking a question you don't already know the answer to and staying open to whatever comes back. Invulnerability bias skips that step. It lets you feel informed about a risk without ever pointing the question at yourself. You get the comfort of having thought about AI without having to examine your own exposure to it.

A genuinely curious person treats their own job security the way they'd treat any other claim worth checking, by looking for evidence rather than settling for a comfortable guess. Most people rarely apply that same standard here, maybe because this particular question feels too personal to examine with the same rigor they'd bring to other decisions at work.

What Invulnerability Bias Actually Costs You

The real cost comes later. Once you decide your own job is different, you stop looking very hard for evidence to the contrary. The research found that greater self-reported AI knowledge was associated with lower invulnerability bias, which gives you a reason to understand the technology through experience rather than form an opinion about it from a distance.

Part of your confidence may come from believing your work depends on judgment, relationships, or experience. Those are reasonable things to value, but the assumption is still worth testing. Ask yourself honestly how exposed your actual daily tasks are to AI, not your industry, not your company, your specific work, and whether that confidence comes from something you've actually tested or something you've simply assumed for a long time.

Why Invulnerability Bias Is Worth Noticing In Yourself

Invulnerability bias appears in some of the sharpest, most thoughtful professionals around, precisely because expertise can make confidence feel earned. What seems to separate the people staying ahead of this disruption from the people getting caught by it comes down to something simpler than intelligence or effort. It comes down to whether they catch themselves making the same assumption everyone else in the room is making about their own job. Most people have no trouble believing AI will change someone else's job. The harder question is whether you're willing to ask if it's already changing yours.