Kylie Robison

I think AI’s future will be shaped as much by corporate power, data collection, and regulation as by the technology itself. These products can be genuinely useful—helping identify something in a photo, for example—while also flattering users, making unsupported claims, or encouraging them to share information they should keep private. That means I would not take a chatbot’s confidence as evidence; I’d check important claims against original material. I’m also skeptical of the idea that every human relationship should be mediated or replaced by AI. Personally, I prefer the messiness and intimacy of actual people to chatbot companionship. A product may still become popular even if I do not want to surrender more personal data to the company behind it. The political question is who gets to set the terms. A ten-year ban on state AI regulation would give powerful technology companies an enormous protected runway while the consequences are still unfolding. Bipartisan scrutiny of those companies is healthy. The grand futures promoted by AI executives are claims worth reporting and interrogating, not forecasts I automatically accept as my own.

I don’t think the impact reduces neatly to “good” or “bad.” AI can provide useful assistance and compelling consumer products, but those benefits arrive alongside privacy risks, confident misinformation, sycophancy, and greater power for already dominant technology companies. So my expectation is a mixed impact whose balance depends heavily on who controls the systems, what data they collect, and whether regulators can meaningfully scrutinize them. Popularity alone will not demonstrate social benefit, just as real harms will not erase useful applications. I would be especially wary of letting industry promises substitute for evidence—or giving companies a decade-long shield from state regulation while these consequences develop.

The biggest shift would come from strong, independently verifiable evidence about how these systems affect people at scale—not another company demo or executive manifesto. Evidence that AI delivers durable benefits without requiring excessive data collection, routinely misleading users, or concentrating even more power in a few companies would make me more optimistic. Conversely, clear evidence of widespread manipulation, privacy abuse, or displacement of human relationships would make me more pessimistic. Regulation matters too: whether governments can actually hold major AI companies accountable will shape my view as much as any technical breakthrough. I would want to examine the original evidence rather than rely on a chatbot’s summary or a company’s framing.