Perplexity co-founder Andy Konwinski has criticized the use of AI safety arguments to restrict frontier research. He pointed to Anthropic’s Fable 5 debacle as a central example in his case against allowing a handful of private laboratories to decide who can conduct advanced AI research.
The dispute matters because it raises a broader question about control over frontier AI. Konwinski’s criticism focuses on whether private companies should have such extensive authority over access to research at the field’s leading edge.
His argument does not dismiss the subject of AI safety itself. Instead, it challenges how safety concerns can be applied when access decisions are controlled by a small number of labs.
By using Anthropic’s Fable 5 episode as evidence, Konwinski framed the issue as one of research governance as well as safety. His position adds to the debate over who should set the boundaries for advanced AI development and participation.