AI for Humanity: What the Phrase Means and Who Is Behind It

“AI for Humanity” isn’t one product or company. It’s a label shared by a few major public-interest AI efforts: France’s national AI strategy (2018), the United Nations report Governing AI for Humanity (2024) and the UN bodies that followed it, and Humanity AI, a $500 million philanthropic fund launched in October 2025. All of them ask the same question: how do we steer AI toward public benefit and away from harms and private capture?

Below we go through each one: what it is, what it has actually done, and where it stands as of October 2026. Several of these programs are still running, so details may change.

Why so many things share one name

The phrase works as a slogan, so it keeps turning up. Besides the three big efforts covered here, you’ll also find it on conference sessions, newsletters and award titles. For example, the AICON 2026 festival in Singapore (8–9 October 2026) opened with a main-stage forum called “AI for Humanity: Technology, Commercialisation and Governance”, according to the event’s announcement.

If you came across the phrase in a news story, the table below should help you work out which effort it meant.

France’s “AI for Humanity” strategy

The phrase first got wide use in March 2018, when President Emmanuel Macron announced a national AI plan called “AI for Humanity”. The plan built on a report by mathematician and MP Cédric Villani, Donner un sens à l’intelligence artificielle (“Giving meaning to artificial intelligence”), which was delivered on 28 March 2018 after extensive hearings and consultations.

The French research ministry lists four priorities:

1. Strengthen the AI ecosystem in France and Europe.
2. Pursue an open data policy.
3. Adapt the regulatory and financial framework.
4. Define the ethical and political stakes of AI.

The European Commission’s AI Watch summary reports that the government committed €1.5 billion to AI through the end of 2022, including €700 million for research. Government summaries credit this first phase with a network of four interdisciplinary AI institutes (the “3IA” institutes), research chairs, doctoral programs and the Jean Zay supercomputer. Those numbers come from the original announcement and later reviews. We have not checked how the figures held up in follow-on phases.

The UN’s “Governing AI for Humanity” report

In September 2024, the UN Secretary-General’s High-level Advisory Body on AI published its final report, Governing AI for Humanity. The report runs about 100 pages and calls for a globally inclusive approach to AI governance, anchored in the UN.

Its main argument concerns who gets a seat at the table. The report found that only seven of the 193 UN member states were party to seven recent prominent AI governance initiatives, and that many countries, mostly in the Global South, were part of none of them.

The key recommendations

The report makes seven recommendations. Coverage from IISD and techUK highlights these:

• An international scientific panel on AI. Independent experts who report regularly on AI capabilities, risks and opportunities.
• A policy dialogue on AI governance. Regular meetings between governments and other stakeholders, rooted in human rights.
• An AI standards exchange. A forum where standards bodies, companies and civil society work on technical interoperability.
• Capacity-building and funding. Measures, including a fund, to help countries that lack AI resources.

To read all seven recommendations in full, use the PDF on the UN Digital Library.

What the UN actually built

The report’s first two ideas became real institutions. On 26 August 2025, the UN General Assembly adopted resolution A/RES/79/325 by consensus. It created an Independent International Scientific Panel on AI and a Global Dialogue on AI Governance, both limited to non-military uses of AI.

Since then:

• February 2026: The General Assembly appointed the panel’s 40 members, drawn from academia, industry, civil society and government. Their three-year term runs to February 2029. One example is Bernhard Schölkopf, who represents Germany. At its first meeting, the panel reportedly elected Yoshua Bengio and Maria Ressa as co-chairs.
• 6–7 July 2026: The first Global Dialogue session took place in Geneva, alongside the ITU’s AI for Good Summit. The panel’s preliminary findings were presented there. Reports quote Bengio as warning that science “currently cannot guarantee” that more capable AI won’t cause catastrophic harm.
• 2027: A second Dialogue session is planned in New York, per the UN panel FAQ.

Sources disagree slightly on when the preliminary report was first published. One secondary summary says 1 July 2026, while UN notices tied its release to the Geneva session.

Humanity AI: the $500 million philanthropy fund

The newest and most hands-on effort is Humanity AI, announced on 14 October 2025. It’s a five-year, $500 million coalition of ten foundations: Doris Duke, Ford, Lumina, Kapor, MacArthur, Mellon, Mozilla, Omidyar Network, Packard and the Siegel Family Endowment (The NonProfit Times). At launch, MacArthur and Omidyar Network were named as co-chairs. Some later coverage also lists Ford as a co-leader.

According to AP’s launch coverage, the goal is to give the public a stronger voice so that AI developers alone don’t decide where the technology goes. The launch named five priority areas: democracy, education, protecting artists, work, and personal security. Reports say Rockefeller Philanthropy Advisors manages the pooled fund.

First grants

On 12 May 2026, Humanity AI announced more than $18 million in grants:

• $8 million to twelve inaugural grantees, including AI Now Institute, Brookings Institution, Data & Society and the Pulitzer Center.
• $10 million held back for a public open call, reported to open in summer 2026 and to prioritize communities closest to AI’s effects.

If you run a nonprofit and want to apply, rely only on the coalition’s official site and the participating foundations’ own pages. Third-party grant listings sometimes show award ranges that the funders haven’t confirmed.

What this means if you use AI tools

None of these efforts gives you a new app or changes how ChatGPT or Claude works tomorrow. They matter in less direct ways:

• Research and evidence. The UN panel’s reports aim to give governments a shared, independent picture of AI risks and capabilities, and that picture can shape future rules.
• Funding for watchdogs. Humanity AI funds groups that study and audit AI, which is where many critical investigations of AI products come from.
• Policy at home. National strategies like France’s decide where research money and computing power go.

If a tool or company markets itself as “AI for humanity”, treat that as a slogan. Judge the product by its privacy policy, its pricing and what it actually does.

How to follow developments

1. For the UN’s work, check the Independent International Scientific Panel on AI pages on un.org and the Global Digital Compact AI page.
2. For Humanity AI grants and open calls, watch the coalition’s site and announcements from members such as MacArthur, Mellon and Omidyar Network.
3. For France, the research ministry publishes updates on the national AI strategy.

As of October 2026, all three are active, and the UN panel and Humanity AI in particular are still producing new reports and grants.