“Anthropic is doubling down on interpretability, and we have a goal of getting to 'interpretability can reliably detect most model problems' by 2027 — the analogue of an MRI for AI, arriving before models become overwhelmingly powerful.”
Co-founder and CEO, Anthropic
Said Apr 2025Deadline Dec 31, 2027Dario Amodei, 'The Urgency of Interpretability', darioamodei.com, April 2025
Stated as an Anthropic goal with an explicit 2027 deadline; resolution will rest largely on Anthropic's own and third-party interpretability evaluations.