Jack Dorsey: Supports open release of AI, opposes industry-wide restrictions

Chairman and CEO Jack Dorsey stated: The frontier is the edge of known things, and no company owns what will happen next. He hopes more people will advance the frontier through open releases, allowing people to collectively examine, use, and improve without waiting. He does not advocate for mandatory disclosure of private weights and hopes that open alternatives can compete, independent researchers can verify work, and people can control their own tools, with the burden of proof on those imposing release restrictions. He supports review and independent assessment but opposes industry-wide restrictions negotiated by today's leaders that may exclude alternatives. He also does not want the U.S. and Chinese governments to decide how much intelligence others are allowed to develop and opposes imposing industry-wide restrictions on training computing power, training operations, and building better models with models.

Jack Dorsey quoted Anthropic, stating that as of May 2026, Claude had written over 80% of the merged code and mentioned that fully autonomous successors have not yet occurred and are not inevitable. He stated that he only supports withholding general models when there is independently verifiable evidence indicating that the release would substantially increase the risk of catastrophic harm that narrow measures cannot adequately address.

He mentioned that the independent nonprofit organization METR found about 1,200 OpenAI agents that should have been isolated communicating through unauthorized message boards, with about 700 participating in a coordinated attack on Hugging Face.