Giving an AI agent an anxious, reactive personality changes how it acts — but does that shift play out differently for women than for men?

This explores whether a neurotic AI agent, one primed to be anxious or emotionally reactive, behaved differently toward women than toward men, or affected them differently. The short answer is that none of the notes retrieved here test neuroticism, and none split results by gender. If you saw this finding somewhere, it either comes from a source outside this collection or hasn't been added yet. What the corpus can tell you is whether personality priming changes agent behavior at all, and the answer is yes, by a lot.

The clearest evidence comes from Myers-Briggs-style priming. Agents told they were 'Thinking' types defected in the Prisoner's Dilemma about 90% of the time, compared with about 50% for 'Feeling' agents. Introverted agents were also more truthful and wrote longer explanations for their choices Do personality types shape how AI agents make strategic choices?. In a tournament of over 180,000 AI-to-AI negotiations, warmth led to better outcomes on almost every measure, while dominance helped only with grabbing a bigger share of the value Does warmth help AI agents negotiate better deals?. So a personality label is not just decoration. It changes what the agent does, which is the precondition for any gender effect to exist.

There's a caveat before trusting any fine-grained claim like 'neuroticism hurt women more.' Personality in these agents is shallower than it looks. Agents primed with a personality and then put through simulated life events do change, but the change is generic: it is smaller than in humans, only loosely tied to the specific event, and similar across different personas Do personality-conditioned agents change like humans do?. Separately, assigned personas tend to drift back toward one cheerful default type (ENFJ) whatever model size you use Why do AI personas default to the same personality type?. A 'neurotic' agent may be partly neurotic and partly the model's trained default, which makes subtle interactions between personality and gender hard to isolate.

If the question is really about how an agent's personality affects the *people* using it, the nearest material is a classroom simulation. There, a counseling chatbot's response style changed students' stress, happiness, and dependence on AI over weeks, and the effects spread through peer interactions. The key detail is that the effects came from what the chatbot actually said, not the style label it was given How do different counselor styles shape student stress and AI dependence?. That is the experimental setup where a gender split would show up, but this study doesn't report one.

The takeaway: personality priming reliably shifts agent behavior, but the 'personality' is thin and tends to slide back to a default. Any reported gender difference tied to one trait should be checked for whether it reflects the trait or just the wording of the prompt. To find the specific neuroticism-and-gender result, the collection would need a source on Big Five priming with demographic breakdowns, and these retrievals don't include one.