Voluntary AI Commitments
This document outlines voluntary commitments made by artificial intelligence companies to manage the risks posed by advanced AI systems. The commitments focus on safety through red-teaming and information sharing, security to protect proprietary model weights and encourage vulnerability reporting, and trust mechanisms like watermarking and public reporting on capabilities and societal risks.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
This document is for AI developers, researchers, and policy makers interested in voluntary corporate commitments and risk management for advanced generative AI systems.
What the document contains
The document covers voluntary commitments made by AI companies in three primary areas: safety (red-teaming and information sharing), security (protecting model weights and incentivizing vulnerability discovery), and trust (watermarking, public reporting, and addressing societal risks). It specifies that commitments apply to frontier models more powerful than current advanced releases.
The document does not provide binding statutory regulations, enforcement mechanisms, or penalties for non-compliance, as these are voluntary industry commitments intended to remain in effect until formal legislation is enacted.
- Companies commit to conducting internal and external red-teaming of models to address misuse, societal risks, and national security concerns.
- Organizations agree to work toward information sharing with governments and industry peers regarding trust and safety risks and safeguards.
- Commitments require investing in cybersecurity and insider threat safeguards to protect unreleased model weights as core intellectual property.
- Companies pledge to develop and deploy mechanisms like watermarking and provenance tools to enable users to identify AI-generated audio and visual content.
- Participating entities agree to publicly report on model capabilities, limitations, and societal risks such as bias and fairness.
· Voluntary AI Commitments · Scope · Safety · Security · Trust
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Accountability
2/5Clear responsibility for AI outcomes
These voluntary commitments also recognize that different AI actors have different responsibilities depending on their roles in the AI development lifecycle
Voluntary AI Commitments
screening score · 2026-10-23
Fairness
3/5Non-discrimination and equitable treatment
Prioritize research on societal risks posed by AI systems, including on avoiding harmful bias and discrimination
7
screening score · 2026-10-23
Human Dignity
0/5Respect for the inherent worth of persons
screening score · 2026-10-23
Human Oversight
0/5Humans remain in control of consequential decisions
screening score · 2026-10-23
Privacy
3/5Protection of personal data and private life
Companies commit generally to empowering trust and safety teams, advancing AI safety research, advancing privacy, protecting children
7
screening score · 2026-10-23
Transparency
4/5Openness about how AI systems work and are used
Publicly report model or system capabilities, limitations, and domains of appropriate and inappropriate use, including discussion of societal risks
6
screening score · 2026-10-23
Enforcement
0/5Consequences and compliance mechanisms with real force
screening score · 2026-10-23
Impact Assessment
4/5Required assessment of an AI system's risks/impacts before or during use
Commit to internal and external red-teaming of models or systems in areas including misuse, societal risks, and national security concerns
Safety
screening score · 2026-10-23