Guidelight's Control Assessment of Frontier AI Companies | Guidelight AI Standards
guidelight.ai · 748 words · saved by 1 readers
Guidelight's August 2026 control assessment of frontier AI companies.
Intro In Guidelight’s first assessment of safety at frontier AI companies, we find that basic practices for keeping control of AI are, at most, partially implemented. Drawing on public materials,1These materials include system cards, safety frameworks, risk reports, blog posts, and third parties' descriptions of collaborations with companies.These materials include system cards, safety frameworks, risk reports, blog posts, and third parties' descriptions of collaborations with companies. we assessed the implementation of Guidelight’s Control standard at five companies: Anthropic, Google,…
saved by
related reading
- gdm-ai-control-roadmap.pdfstorage.googleapis.com
- We need 3rd party Training-Run Assessments — LessWronglesswrong.com
- Frontier Risk Report (February to March 2026) - METRmetr.org
- The case for ensuring that powerful AIs are controlled — LessWronglesswrong.com
- Common Elements of Frontier AI Safety Policiesmetr.org
- Towards Best Practices in AGI Safety and Governance | GovAIgovernance.ai
- Standards Development Process | Guidelightguidelight.ai
- AI Control: Improving Safety Despite Intentional Subversion — LessWronglesswrong.com
- metr.org/risk-report-feb-mar-2026.pdf#page=33.47metr.org
- The Case Against AI Control Research — LessWronglesswrong.com
- How our new Control Red Team is stress-testing frontier monitors | AISI Workaisi.gov.uk
- Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com