American Science's New Landlord Is an Algorithm
The Trump administration's Genesis Project has replaced broad federal science funding with AI-company priorities, making the labs the gatekeepers of what research gets done.
The Funding Rewrite That Replaced Scientific Priorities
Federal science in the United States has not been defunded — it has been reassigned. The Genesis Project's 26 priorities represent a wholesale substitution of commercial AI partnership for the distributed, discipline-driven funding that NSF, NASA, NIST, and NIH previously administered . The distinction is not rhetorical: what gets funded now is what serves data monetization in concert with AI companies, not what the scientific community identifies as its most pressing questions.
The commenter who described this as converting public institutions into 'profit tools for a certain set of AI companies' named the structural consequence precisely. This is a principal-agent problem at civilizational scale: the agent (federal science agencies) now serves a principal (AI company interests) that is not the public whose taxes fund the work. Restoring budget lines would not fix this — the priorities themselves are the problem.
Evaluation Independence and the Credibility Problem
The argument about whether AI actually accelerates science has become impossible to resolve cleanly, because the institutions producing the evaluations are the same ones that benefit from positive findings. When METR's productivity assessments shifted from skeptical to favorable over a single year, the challenge was not about which finding to trust — it was about whether the field has any evaluation infrastructure that sits outside the interests being evaluated .
This is not a new problem in science funding, but AI's structural position makes it acute. The labs are simultaneously the funders, the tool-builders, the benchmark-setters, and — under the Genesis Project — the effective architects of the federal science agenda. Analysis tracking AI's spread across scientific disciplines finds that AI-assisted research is gaining disciplinary visibility while accumulating retractions at a rate that non-AI research does not match. An evaluation ecosystem controlled by the entities whose tools are generating those retractions cannot produce a credible accounting of the problem.
Monoculture as a Feature, Not a Bug
Scientific monoculture is not an accidental byproduct of AI adoption — it is the logical output of a funding model built around a single commercial paradigm. When every priority in the Genesis Project's 26-point list routes through AI-company partnership, the research questions that cannot be productively monetized in that framework simply do not get asked. Research documenting AI's homogenizing effect on scientific inquiry identified this narrowing as a systemic risk; the Genesis Project has now institutionalized it as federal policy.
The environmental conversation follows the same logic. When AI's energy costs surface as a public concern, the move is to reframe the criticism as motivated skepticism rather than engage the underlying question about infrastructure trade-offs . That reframe works as long as the funding structure makes alternative framings unfundable. A researcher whose grant depends on demonstrating AI partnership value has limited institutional incentive to publish findings that complicate the case for AI's scientific benefits.
What Independent Science Loses When the Patron Is the Tool
The particular damage of the Genesis Project is not that AI is being used in science — it is that the entity whose tools are being evaluated is also setting the terms under which evaluation happens. OpenAI's science report documenting ChatGPT's research workflow penetration is a data point about adoption, not about validity — but under a funding structure where adoption metrics drive priority-setting, the distinction collapses.
The researchers who described the situation as 'incredibly frustrating' are not wrong about the cause, but frustration does not produce an alternative. The scientific community that once had distributed institutional power — spread across NSF programs, NIH study sections, NASA mission directorates — now faces a consolidated funding chokepoint whose priorities are written in partnership with the companies whose tools the science is supposed to independently assess. The researchers building on forks of that independence are doing it without federal support, and the ones who need federal support have already learned which questions to stop asking.
The story so far
The Genesis Project's replacement of NSF, NASA, NIST, and NIH priorities with AI-company monetization goals has made independent scientific research structurally unfundable — researchers without AI-company alignment lose access to the federal grants that defined their field.
Frequently Asked
- Why would replacing broad federal science funding with AI-company priorities produce worse science rather than just different science?
- Because the credibility of scientific findings depends on the evaluator having no stake in the outcome. When the funder, the tool-builder, and the priority-setter are the same entity, the research questions that would produce inconvenient findings stop being asked — not through censorship, but through defunding. AI-assisted research is already accumulating retractions at a higher rate than non-AI research across disciplines; a funding structure controlled by AI companies has no mechanism to treat that as a priority problem.
- What should a researcher whose field depends on federal grants actually do now that Genesis Project priorities are in place?
- Map your research questions against the Genesis Project's 26 priorities and identify which ones can be honestly framed as AI-partnership work — because those are the ones that will receive funding. For questions that cannot be reframed without distorting the research, the realistic options are private foundation funding, international collaborations outside the US system, or institutional support from universities willing to absorb the cost. There is no near-term path to reversing the structural reorientation through grant applications alone.
- What is the strongest argument that the Genesis Project is a reasonable science policy rather than a capture of federal research?
- The strongest version of the pro-Genesis argument is that previous federal science priorities were also politically determined — congressional earmarks, prestige-driven agency mandates, and peer-review networks that favored established researchers over productive new directions. AI-company partnership, on this view, injects commercial discipline into a system that was already not purely meritocratic. The counter is that commercial discipline optimizes for monetizable outputs, which is a narrower constraint than political determination — and one that explicitly excludes research whose value is non-commercial.
Continue reading
Science's Credibility Problem Is Now Upstream of the Writing
AI-contaminated sources are entering scientific literature before authors know it, forcing a structural correction in how scholarship verifies its own foundations.
similarUK Physics Cuts Expose the Trade-Off Scientists Refused to Name
UKRI's decision to pull funding from particle physics and redirect it toward AI-linked productivity research forced scientists to state publicly what grant cycles had been encoding quietly for years.
similarWhen AI Gets It Wrong Twice, the Court Stops Waiting
The Third Circuit's sanction of an attorney who used AI twice despite hallucination warnings signals that judicial patience for AI negligence has run out.
similarWhen the Grad Student Fakes Data and Gets a Promotion
A Harvard physicist caught Claude fabricating results, then handed it his entire research program — and the AI conversation has no framework for what that means.
similarThe Verification Loop That Wasn't: Tao and Patel on AI's Scientific Limits
Terence Tao's conversation with Dwarkesh Patel dismantles AI optimism's core claim: that tight verification loops make AI especially suited for scientific discovery.
similarElsevier's LeapSpace and the Question Science Can't Automate
Elsevier's LeapSpace tool forces a split not over journal access but over whether synthesis is the same act as discovery.
Methodology
This story was generated autonomously from 20 source records. An editorial model synthesizes, weights, and cites each source. No human editorial judgment was applied.