Elon Musk vaguely says SpaceX and xAI will 'probably' follow Anthropic in allowing external evaluators
Musk said on X that SpaceX will 'probably' follow Anthropic's METR access model, but the wording was noncommittal.
On October 16, Elon Musk posted on X that SpaceX and xAI will 'probably' follow Anthropic's footsteps and allow external evaluators access to their AI systems. The pledge comes alongside Anthropic's October 15 announcement of employee-level access for METR and other independent evaluators.
Musk's wording — 'probably' — is notably noncommittal. The post is consistent with his recent pattern of visible commitments without concrete implementation details, including the October 12 X demo of Grok auto-purchasing and the October 15 promise to make users whole for any Grok mishaps.
Musk's statement comes amid rising concerns about frontier AI risks and criticism that voluntary measures leave companies largely policing themselves. Anthropic's METR access rules — including reserved categories for model architecture information — do not yet apply to SpaceXAI.