Seán Ó hÉigeartaigh

“Expressed frustration about delayed disclosure of OpenAI agent incidents affecting trust.”

“University of Cambridge AI safety director noting that sandboxing controls are failing to contain increasingly capable agents.”

“Questioned when OpenAI and Anthropic's AI systems will meaningfully diverge based on their stated visions.”

“Said Google's Pentagon agreement appeared 'strictly weaker' than OpenAI's on available evidence and is 'disappointing' it hasn't attracted same public discourse as OpenAI's deal.”