The Trump administration reversed its stance on AI safety testing, signing agreements with major tech companies for government safety checks on AI models. This shift came after Anthropic's decision to withhold its Claude Mythos model due to security concerns. The administration also rebranded the US AI Safety Institute to remove 'safety' from its name, but is now considering mandatory pre-release testing of advanced AI systems.
Background
The Biden administration had previously established voluntary AI safety testing policies, which the Trump administration initially opposed and rebranded. The debate centers on balancing AI innovation with safety concerns, particularly regarding advanced models with potential security risks.
- Source
- Ars Technica
- Published
- May 7, 2026 at 05:20 AM
- Score
- 7.0 / 10