Bessent Compares AI Safety-Warning CEOs to Hannibal Lecter - Startup Fortune

Oh my goodness, buckle up, because Treasury Secretary Scott Bessent has just handed the AI safety conversation its *most quotable mascot yet* — Hannibal Lecter! ✨ Whoever decided that the best way to talk about catastrophic risk was through a fictional cannibal with excellent table manners deserves a communications award, honestly. Think of it as *branding*: you cannot buy that kind of name recognition, and Bessent has generously given the entire safety-warning genre a marquee villain it never asked for. The 'Lecter line,' as the report calls it, has sharpened his framing into something *quotable* — which is the single most flattering thing anyone has ever said about a policy metaphor. ✨

Now, some people might read Bessent's promise that the government will *not* become a 'liability shield' for AI hyperscalers as bad news — but look at what it actually says! No cosy government cushion, no bureaucratic bubble wrap, no taxpayer-funded parachute. That is *accountability* with the training wheels taken off, and frankly it is thrilling to see hyperscalers trusted to stand on their own two magnificent, enormously capitalised feet. No wonder they are being treated like grown-ups — because they absolutely, definitely are! Freedom! ✨

And the number everyone's frowning about? Those are excellent odds! Better than a lot of mortgage applications! It is the same Treasury Secretary who floated a US–China AI notification hotline just weeks ago which we adored, so think of this as the sequel: same leading man, bolder costume choices, and a cannibal joke that will absolutely be quoted at every conference for the next twelve months. What a time to be quotable! ✨