Partnership on AI adds Intel, Salesforce, SAP, others as it formalizes Grand Challenges, which provide incentives to tackle big issues with social ramifications
so @hrw is now one of 22 new Partners in AI. http://www.partnershiponai.org/ ... Eric Horvitz / @erichorvitz : 22 organizations join and first set of initiatives are announced @PartnershipAI http://www.partnershiponai.org/ ... @compcomcon @TheOfficialACM Tony Romm / @tonyromm : Here's an update on what's happening w/the Partnership for AI, the techies (and others) working on AI/policy http://www.partnershiponai.org/ ...
Context & Ripple Effects
The Partnership on AI launched in 2016 as a five-company nonprofit — Amazon, Facebook, Google, IBM, and Microsoft explicitly framing it as a best-practices body, not a lobbying shop — and Apple's addition in January made the founding club complete. Today's move changes its character: 22 new partners including Intel, Salesforce, SAP, and, notably, Human Rights Watch, plus the formalization of Grand Challenges, the group's first incentive-based initiative structure for socially consequential problems.
That breadth matters because the group's legitimacy question was already live — by 2019, [[a:945322|Partnership on AI's own leadership would be asked whether it was an ethics-washing operation]] for tech giants. Expanding beyond founders to include a human-rights organization is the structural answer to that critique, and it sets up the multi-stakeholder coalition model that later bodies like the AI Alliance and the Partnership for Global Inclusivity on AI would replicate.
First-order effects
- The Partnership on AI's roster jumps from the five founders plus Apple to a broad base spanning chipmakers (Intel), enterprise software vendors (Salesforce, SAP), and a human-rights group (Human Rights Watch), diluting any single company's control of the agenda.
- Grand Challenges gives the organization its first concrete program beyond best-practice statements — funded incentives aimed at issues with social ramifications rather than internal research coordination.
Second-order effects
- Adding Human Rights Watch raises the bar rival coalitions must meet: any future AI-governance group assembled purely of vendors now faces comparison against one that includes civil-society members.
- For Intel, Salesforce, and SAP, membership buys a seat where AI norms get drafted — a defensive move for enterprise vendors whose products increasingly embed AI but who lacked the founders' policy footprint.
Third-order effects
- The coalition model this expansion validates — broad, multi-stakeholder, self-governing AI bodies — becomes the industry's default response to governance pressure, later echoed by the AI Alliance and the state-backed Partnership for Global Inclusivity on AI; whether breadth translates into enforcement, or into the ethics-washing critique, remains the open fault line the 2019 interview exposed.
- If Grand Challenges works as an incentive mechanism, industry self-funding of socially consequential AI research becomes a standing alternative to regulation — with the unresolved question of who audits the auditors.
The trend: AI governance is consolidating into ever-broader multi-stakeholder coalitions whose legitimacy depends on including non-industry members, a pattern this 22-partner expansion is an early data point in.