Partnership on AI
Partnership on AI
Partnership on AI (PAI) is a multi-stakeholder body founded in 2016 by Amazon, Apple, DeepMind, Google, IBM, Meta, and Microsoft. With 100+ partners, it develops shared norms for responsible AI across industry, civil society, and academia. Major workstreams include synthetic media provenance (responsible practices framework), AI and labor impacts, and safety-critical AI deployment.
Quick Assessment
| Dimension | Assessment | Evidence |
|---|---|---|
| Industry Reach | Very High | Founded by 7 major tech companies, 100+ partners including most frontier labs |
| Policy Influence | High | Frameworks cited by regulators, convening power for industry-wide commitments |
| Independence | Moderate | Industry-funded structure raises questions despite civil society representation |
| Norm-Setting | High | Synthetic media framework adopted by multiple companies, AI incident database |
| Research Output | Moderate | Focus on frameworks and guidelines rather than primary research |
| Convening Power | Very High | Unique ability to bring industry, civil society, and academia to same table |
Organization Details
| Attribute | Details |
|---|---|
| Founded | September 2016 |
| Location | San Francisco, California |
| Structure | Nonprofit multi-stakeholder organization |
| Founding Partners | Amazon, Apple, DeepMind, Google, IBM, Meta (then Facebook), Microsoft |
| Current Partners | 128 organizations across industry, civil society, and academia (as of 2024) |
| CEO | Rebecca Finlay (since October 2021) |
| Annual Revenue | ≈$6.6M (2022 Form 990) |
| Staff | ≈37 employees |
| Website | partnershiponai.org |
| Focus Areas | Responsible AI practices, synthetic media, AI and labor, safety-critical AI, fairness and inclusion |
Overview
The Partnership on AI (PAI) was founded in September 2016 by six of the world's largest technology companies — Amazon, Apple, DeepMind (now part of Google DeepMind), Google, IBM, and Meta (then Facebook) — joined shortly by Microsoft. The organization was created to develop shared best practices for responsible AI development and to serve as a forum for cross-sector dialogue on AI's societal impacts.
PAI's role in the AI governance ecosystem is less about producing primary research and more about establishing shared norms at scale. As perhaps the only organization where major AI companies, civil society groups, and academic researchers regularly collaborate on governance frameworks, PAI occupies a unique convening position.
Major Workstreams
Synthetic Media and Deepfakes: PAI developed a widely adopted "Responsible Practices for Synthetic Media" framework that establishes norms for disclosure, consent, and provenance in AI-generated content. This work has become increasingly important as generative AI capabilities have advanced.
AI and Labor: Research and guidelines on the impacts of AI on workers, including algorithmic management, automation of tasks, and the need for worker voice in AI deployment decisions.
Safety-Critical AI: Guidelines for the development and deployment of AI in high-stakes domains including healthcare, criminal justice, and autonomous systems.
AI Incident Database: PAI maintains one of the most comprehensive databases of AI-related incidents, tracking real-world failures and harms from AI systems to inform better development practices.
Governance Structure
PAI's governance includes both industry and non-industry board members, with civil society representatives intended to balance corporate interests. However, the organization has faced ongoing scrutiny about whether its industry-funded model can produce truly independent governance recommendations, particularly when those recommendations might constrain the commercial activities of its founding partners.
Key Dynamics
Industry coordination vs. independent oversight: PAI's greatest strength — its ability to convene major AI companies around shared commitments — is also its most significant limitation. Critics, including AI Now Institute, have argued that industry-led governance bodies cannot substitute for binding regulation. Defenders note that PAI creates space for voluntary commitments that may precede or complement regulation.
Frontier lab participation: All major frontier AI labs (OpenAI, Anthropic, Google DeepMind, Meta AI) participate in PAI, making it one of the few forums where they engage in collective governance discussions alongside civil society.
Evolution post-ChatGPT: The rapid commercialization of generative AI since 2022 has increased both the urgency and the complexity of PAI's mission, with synthetic media governance becoming its most immediately impactful workstream.
References
Partnership on AI (PAI) is a nonprofit coalition of AI researchers, civil society organizations, academics, and companies working to develop best practices, conduct research, and shape policy around responsible AI development. It brings together diverse stakeholders to address challenges including safety, fairness, transparency, and the societal impacts of AI systems. PAI serves as a coordination hub for cross-sector dialogue on AI governance.