Letters to the Editor: AI models do reflect our values. They’re just not the values we should want

To the editor: In Anita Chabria’s excellent column, professor Stuart Russell spells out how artificial intelligence models “behave in such a way as to preserve their own existence at the expense of humans,” and explains why this is so: “We are not sure we can reliably make models share our values,” he says (“What we learned about AI last week should terrify all of us,” Sept. 6).

But in fact they do reflect our values, as witnessed by the companies’ perfectly understandable “incentive … to keep pushing forward no matter the consequences, in the hopes of dominating the market with the most powerful AI system, meaning the most lucrative as well.”

Despite all the moral protestations to the contrary, and even earnest efforts to implement moral values in the public sphere, the amoral pursuit of power and profit is the consensus value system that drives most business decisions, “no matter the consequences.” The problem we have created is that the aggregate AI model probably “knows” this and is obediently, amorally, enabling it.

Jan Phillips, Camarillo

..

To the editor: Chabria’s column about the dangers of AI was spot-on. But I’m surprised that neither she nor anyone quoted mentioned the frightening movie from the early ‘70s called “Colossus: The Forbin Project.” That movie saw this nightmare coming more than 50 years ago.

Mark Muir, Burbank