To explore how AI is understood and experienced across different corners of campus, I interviewed three people. A fellow undergraduate studying data science, a professor in the School of Information, and a product manager who recruits Berkeley students and spoke with me over a coffee chat. Their answers surprised me in how much they diverged and in a few places, unexpectedly aligned.
My peer was candid: she thinks the biggest risk isn’t rogue superintelligence but mundane bias baked into training data. She emphasized the need for diverse engineering teams as a structural safeguard not just a values statement. The professor took a more policy-oriented view, arguing that responsible AI requires external accountability that self-regulation by labs has already proven insufficient, and that something like the FDA model for high-stakes AI systems deserves serious consideration. The PM, coming from industry, was more pragmatic: she believes “responsible AI” often becomes a PR exercise unless companies tie safety benchmarks directly to shipping decisions, which almost none currently do.
The data science student described a shift in what’s actually valued in her coursework less emphasis on writing models from scratch, more on knowing how to evaluate, fine-tune, and critically interrogate outputs. She said it’s made her feel both more capable and more uncertain about what her skills are actually worth. The professor noted that AI has already restructured how knowledge work gets done in research contexts. The PM said AI has quietly transformed recruiting: résumé screening is nearly fully automated at her company now, which she acknowledged creates real fairness problems they haven’t solved.
What struck me most was how much each person’s position shaped their answer not just their expertise, but their relationship to risk, to institutions, and to their own career
Related Challenge:
challenge_points:
Image:
