Future of Life Institute grades 9 AI labs on safety
OpenAI, Anthropic, and 7 others ranked in Summer 2026 AI Safety Index
The Future of Life Institute (FLI) has published its Summer 2026 AI Safety Index, a first-of-its-kind comparative assessment grading nine prominent AI labs—OpenAI, Anthropic, Google DeepMind, xAI, Zhipu AI, Meta, DeepSeek, Alibaba Cloud, and Mistral AI—on critical safety and security domains.
The index, developed with input from independent experts, evaluates each lab across key risk management metrics, including model transparency, red teaming practices, and alignment research. This initiative aims to provide stakeholders—from policymakers to investors—with a transparent framework to assess how frontier AI developers are mitigating risks like misuse, misalignment, and systemic failures.
- First-ever comparative AI safety index grades 9 labs: OpenAI, Anthropic, Google DeepMind, xAI, Zhipu AI, Meta, DeepSeek, Alibaba Cloud, and Mistral AI
- Evaluates key safety domains: model transparency, red teaming, alignment research, and risk mitigation
- Provides stakeholders with a transparent benchmark to assess frontier AI labs' safety practices
Why It Matters
Sets a precedent for standardized AI safety benchmarking, shaping industry accountability and policy.