Why AI Makes Us So Angry. And why the rage depends more on your… | by Livestock Dev | Jun, 2026 | MediumSitemapOpen in appSign up<br>Sign in

Medium Logo

Get app<br>Write

Search

Sign up<br>Sign in

Why AI Makes Us So Angry

Livestock Dev

2 min read·<br>Just now

Listen

Share

And why the rage depends more on your day than on the machine<br>Anyone who works with AI has felt it: a flash of rage that seems out of proportion to a chatbot getting something wrong. What’s interesting is that the intensity of that anger often tracks our own state more than the AI’s actual behavior.<br>A few mechanisms stack on top of each other. AI is fluent and confident, which sets an expectation of competence — so when it misreads us, ignores a constraint, or states something wrong with total assurance, the gap between how capable it sounded and how it performed is jarring. Expectation violation is a core anger trigger. And because the interface is language, the brain routes the interaction through its social systems: a failure feels less like a tool breaking and more like a person refusing to listen, which is far more infuriating. It’s also a black box. With an ordinary tool we can usually see why it failed and adjust; here we can’t form a clear fix, so repeated failures produce a helpless, looping frustration. The effort is asymmetric, too — you explain, it drops half of what you said, you re-explain, it breaks something else — and each cycle feels like wasted effort.<br>But the most telling part is how the intensity tracks our own state. Frustration tolerance isn’t fixed; it’s a resource that depends on physiology and mood. When we’re tired, hungry, stressed, or already emotionally activated, our capacity for top-down regulation drops, so the same trigger crosses the anger threshold that it wouldn’t have on a good day. It’s less that the AI changed and more that the buffer shrank. Two things compound this: working with AI is cognitively taxing — holding context, judging output, correcting it — and that load is most aversive exactly when reserves are low; and in a negative state we read ambiguous things more hostilely, so an imperfect-but-neutral response gets interpreted as the machine being obtuse. Then a feedback loop kicks in: anger is itself depleting, which lowers the threshold further, making the next glitch land harder.<br>The takeaway is almost reassuring. The rage isn’t really caused by the AI. The AI’s friction is fairly constant — what changes is how much of it we can absorb before tipping over. On a bad day, the machine didn’t get worse. Our tolerance just got thinner.

This is an observation and a hypothesis, not a literature review. It draws on well-established ideas about expectation violation, cognitive load, and frustration tolerance, but the specific way they combine here is my own read — offered as a lens, not a citation.<br>Press enter or click to view image in full size

Rage Against The Machine

Claude Code

AI

Psychology

Sociology

Written by Livestock Dev<br>0 followers<br>·1 following

Help

Status

About

Careers

Press

Blog

Privacy

Rules

Terms

Text to speech