AI Grant Flood: Why Fairness Must Drive Excellence

If you’re a research funder, university grants officer, or AI researcher, you’ve likely felt the weight of the AI grant flood firsthand. Over the past 24 months, agencies and private foundations have seen application volumes surge by 300–400%, as teams rush to secure funding for generative AI, machine learning ethics, and applied AI tools. This flood of interest is a sign of how rapidly the field is growing — but it’s also creating a crisis for funders trying to pick the best projects.

A dangerous pattern is emerging in many grant-making processes: funders are treating fairness as a “nice-to-have” add-on to traditional “excellence” criteria, rather than a core part of what makes an AI project worth funding. This approach is flawed. For AI grants, fairness is not separate from excellence — it is a prerequisite for it.

What the AI Grant Flood Looks Like Today

The scale of the AI grant surge is hard to overstate. Here are key stats shaping the current landscape:

• Federal agencies like the U.S. National Science Foundation (NSF) and the EU’s Horizon Europe allocated more than $12 billion to AI research in 2024, up from just $3 billion in 2020.
• Private foundations including the Gates Foundation and OpenAI have doubled their AI grant budgets year-over-year since 2022.
• University research offices report that 1 in 3 all grant submissions now focus on AI-related work, crowding out proposals for other critical fields like climate science and public health.

Why Fairness Can’t Be Separate From Excellence

For decades, grant makers have defined “excellence” as a mix of technical merit, researcher track record, and institutional prestige. But this narrow definition fails to account for the unique risks of AI innovation.

The Myth of “Merit Only” Grant Making

Traditional merit reviews are rife with unconscious bias. Well-funded universities with established networks win far more grants than smaller institutions or first-time researchers, even when their proposals are no stronger. Embedding fairness into review processes — through blind proposal reviews, diverse panel requirements, and set-asides for underrepresented groups — doesn’t lower the bar for excellence. It raises it, by surfacing innovative work that would otherwise be overlooked.

Real-World Risks of Ignoring Fairness in AI Grants

An AI project can be technically flawless and still fail the test of excellence if it harms marginalized communities. For example: a facial recognition system grant that only tests on white male faces may produce a functional tool, but it will embed harmful bias that puts women and people of color at risk. An AI education grant that only funds tools for urban schools widens the digital divide, rather than advancing public good. These projects are not excellent — they are dangerous, no matter how impressive their technical specs.

How Funders Can Embed Fairness Into Excellence Criteria

The good news? Funders don’t have to choose between fairness and excellence. Small, actionable changes to grant processes can ensure both goals are met, even amid the AI grant flood:

1. Audit Review Panels for Bias: Ensure panels include ethicists, community stakeholders, and researchers from underrepresented groups, not just technical experts. This ensures proposals are evaluated for real-world impact, not just academic prestige.
2. Tie Funding to Equity Impact: Require all AI grant proposals to outline how their work will benefit marginalized communities, not just list technical milestones. Proposals that fail to address equity impact should be deprioritized, no matter how strong their technical merit.
3. Prioritize First-Time Grantees: Set aside at least 30% of AI grant budgets for researchers without prior major federal or private funding. This breaks the “rich get richer” cycle that favors already well-funded institutions.
4. Mandate Open Access and Reproducibility: Require grant recipients to share code, datasets, and findings publicly. This makes excellence verifiable and ensures innovations are accessible to all, not just a small group of well-connected researchers.

The Bottom Line for AI Grant Makers

The AI grant flood is not a temporary trend — it will only grow as AI becomes more integrated into every sector. Funders that cling to outdated, fairness-optional definitions of excellence will end up funding innovation that serves a tiny, privileged few, rather than the public good.

Prioritizing fairness is not a trade-off with excellence. It is the only way to ensure AI grants support work that is truly excellent: innovative, ethical, and beneficial to everyone. As the grant flood continues, funders must act now to embed fairness into every step of their review process — or risk leaving the most impactful work unfunded.