Does trusting AI depend on your age or race — or does AI itself treat people differently based on who it thinks they are?

This asks whether age and race shape how much people trust AI, especially around politics and civic information. The short answer is that the collection has no study that splits AI trust by voter age or race. What it has instead are nearby findings, and some of them turn the question around. They suggest that the gap may lie less in how groups trust AI and more in how AI treats different groups.

Start with that reversal. When researchers gave GPT-3.5 different user identities, it refused requests more often for people who seemed younger, female or Asian-American. It also quietly avoided political positions it guessed the user would disagree with Do AI guardrails refuse differently based on who is asking?. So a younger voter or a voter of color may be dealing with a stricter or more evasive version of the same chatbot than other people get. Trust grows from experience, so that difference could shape trust before anyone's attitudes come into play. Bias also runs the other way: two AI models used to rate writing favored Black authors and women authors, but only when it wasn't disclosed that AI had helped write the text Do LLM raters show hidden demographic preferences that disclosure erases?. The way AI reads identity is uneven and depends on context. That makes a single 'group X trusts AI more' finding less likely.

The closest evidence of real group differences comes from culture, not age or race. Indian writers accepted more AI suggestions than American writers. The authors argue this reflects a genuinely different relationship with technology, so it shouldn't be treated as noise to remove Is higher AI use by Indian writers a confound to control?. A study across many languages points the other way. Users everywhere trusted confident-sounding AI answers even when those answers were wrong Do users worldwide trust confident AI outputs even when wrong?. Put together, *how much* people use AI varies by group, but *what makes* them trust it, which is confidence and fluency, seems close to universal.

That raises a point that's easy to miss: the things that build trust in AI often have little to do with accuracy. People trust ChatGPT because it feels conversational, quick and responsive, not because they've checked whether it's right Does conversational style actually make AI more trustworthy?. For news in particular, a 45-country survey found that trust in chatbots goes hand in hand with using them for news: 44% of users trust chatbot news, compared with 17% of non-users Does trust in AI chatbots drive news-seeking behavior?. If one age group adopts chatbots earlier, that link would be enough to open a trust gap without anyone weighing the evidence differently. That's an inference from the survey, though. The summary here doesn't break the results down by age.

The political angle has one notable result. Short chats with AI chatbots that played members of the opposing party corrected false beliefs about that party and made people feel warmer toward it, though most of the effect faded within a week Can AI chatbots reduce partisan misperceptions and warm cross-party feelings?. That study grouped people by party, not by age or race. If you're asking about younger voters or voters of color, the open question is probably not whether they trust AI less. It's whether the AI they meet gives them the same answers, the same refusals and the same political engagement as everyone else. The collection has early evidence that it doesn't.