Epstein-backed funding linked to AI scholar's HK$8.9 million in Hong Kong grants - what public officials should take from this
US government records indicate Jeffrey Epstein committed at least US$113,000 to American researcher Ben Goertzel, known for popularising the term "artificial general intelligence." That support was connected to at least HK$8.9 million in Hong Kong government grants awarded through a top city university.
Goertzel previously led AI and robotics initiatives in Hong Kong. The funding trail raises questions about disclosure, due diligence, and how external backers intersect with public research money.
What the records suggest
Private funds from Epstein were used to back Goertzel's work and, according to the documents, helped him secure substantial local government grants. The case spotlights a practical issue for grant-making bodies: how to assess external sponsorships that sit adjacent to, or directly support, public funding applications.
Put plainly: even modest third-party contributions can influence eligibility perceptions, review outcomes, or reputational risk after the fact.
Why this matters for public funding integrity
- Undeclared or opaque private backing can bias selection processes or erode trust in award outcomes.
- Reputational spillover can trigger audits, delays, or clawbacks-consuming staff time and budget.
- Cross-border donors add complexity for verification, sanctions checks, and conflict-of-interest screening.
Immediate actions for grant administrators
- Tighten disclosure forms: Require full accounting of all external cash and in-kind support tied to the proposal, lab, or principal investigator for the past 3-5 years.
- Source-of-funds verification: Mandate attestations and documentation when private donors are involved; perform sanctions, PEP, and adverse-media checks.
- Conflict-of-interest attestation: Expand COI to include donors, foundations, and intermediaries-not just corporate sponsors.
- Independent review layer: Route proposals with external backers to a secondary integrity review before final award.
- Clawback and suspension clauses: Bake in triggers tied to undisclosed funding, misstatements, or material reputational events.
- Public disclosure registry: Publish non-sensitive sponsor information post-award to reinforce transparency.
- Ongoing monitoring: Re-screen awardees and known sponsors at key milestones and before tranche releases.
Policy updates to consider this quarter
- Eligibility rules: Disallow applications where any associated donor fails basic screening or refuses reasonable transparency.
- Donor intermediaries: Treat gifts via foundations or consultancies as high-risk until fully documented.
- Grant scoring: Introduce a small negative weighting for proposals with unverifiable private funding sources.
- Audit trail: Standardize a short "integrity memo" for each award, capturing checks performed and findings.
Handling AI research proposals with external backers
- Segregate funds: Require separate accounts for public and private monies with clear cost-allocation rules.
- Data and IP controls: Ensure private sponsors get no preferential access to datasets, compute, or IP beyond approved agreements.
- Publication transparency: Mandate sponsor disclosures in papers, demos, and press statements tied to grant-funded outputs.
- Safety review: For AI/robotics, add a lightweight safety and misuse assessment to the integrity screen.
What to watch next
- Any institutional reviews or audits connected to the grants in question.
- Updates from authorities tied to financial crime, due diligence standards, or cross-border donor scrutiny.
- Whether universities tighten their own gift acceptance and disclosure policies for research units.
For context on Hong Kong's public research support mechanisms, review the Innovation and Technology Fund's official guidance here. Background on Epstein's criminal case is available from the US Department of Justice here.
Upskilling your team
If you're updating grant oversight or AI review workflows, a quick way to close skill gaps is targeted training on AI methods, governance, and audit practices. See role-based options here.
Your membership also unlocks:
