The Partnership on Artificial Intelligence to Benefit People and Society, known as Partnership on AI or PAI, was announced on 28 September 2016 by five of the largest technology companies working on AI: Amazon, Facebook, Google together with DeepMind, IBM, and Microsoft. Leading researchers from those companies - among them Eric Horvitz of Microsoft, Francesca Rossi of IBM, Yann LeCun of Facebook, Mustafa Suleyman of DeepMind, and Ralf Herbrich of Amazon - convened to create a shared forum for the field’s hardest ethical questions.
PAI was set up as a non-profit that deliberately did not give the founding corporations exclusive control. Its structure was designed so that academic, civil-society, and other non-corporate organizations could sit in leadership alongside the big labs, and its membership grew to span industry, academia, civil society, and media. The organization describes its mission as bringing “diverse voices together across global sectors, disciplines, and demographics so developments in AI advance positive outcomes for people and society.”
Rather than write binding rules, PAI focuses on producing best practices, research, and shared resources. Its early thematic pillars covered ethics, fairness and inclusivity; transparency, privacy and interoperability; collaboration between people and AI systems; and the trustworthiness and robustness of the technology. Among its most cited contributions is the AI Incident Database, launched in 2020 to catalog real-world AI harms.
Why business readers should care: Partnership on AI showed that competitors could cooperate on safety and ethics norms without a regulator forcing them, and the practices and incident data it produces are widely used as informal standards across the industry.