At a recent AI conference, Garrison Lovely said the most unsettling thing he heard wasn’t a prediction but the mood in the room: a senior person at one of the leading AI companies told him he expects advanced AI to concentrate enormous power either inside the companies building it or in the hands of political leaders who control it. That person, Lovely noted, was not an “AI safety” specialist but someone working on capabilities.
Lovely’s larger takeaway was stark: “There are no adults in the room.” He added that even people working directly on alignment were admitting they do not have a clear solution for how to control systems far smarter than us.
Lovely said he came away convinced the fear inside these companies is real — not marketing, not hype — fear from the people actually building the thing.
Advanced AI concern among builders
The account emphasizes that the worry expressed at the conference came from engineers focused on capabilities rather than safety, and that alignment researchers acknowledged an absence of clear control strategies for much-more-intelligent systems. Lovely framed the overall atmosphere as one of genuine fear among those developing advanced AI.
