The Epstein files.
The unanswered connection.

The public record establishes that Jeffrey Epstein supplied money to scientific and university institutions, made repeated visits and maintained relationships with researchers. It does not follow from those facts that the recipients were carrying out clandestine behavioral experiments, receiving money for the CIA, or continuing MKULTRA.

In January 2020, MIT published the results of an independent investigation into its engagements with Epstein. It reported ten gifts totaling $850,000 between 2002 and 2017, including support directed to the Media Lab and to research associated with Seth Lloyd. The review examined who accepted the money, the approvals involved and Epstein's visits. These findings provide transactions and named institutional decisions that can be checked in the underlying report, with the unusual post-conviction approval process treated as its own historical subject. Original MIT investigation EP-005 ↗

Harvard's institutional review separately reported $9,179,000 in Epstein gifts before 2008 and identified a $6.5 million gift in 2003 to establish the Program for Evolutionary Dynamics. The report also described Epstein's continuing relationship with the program and his visits after his imprisonment. Its findings support precise analysis of funding and institutional access, while the report's scope and the documentation it examined do not establish a link to any CIA behavioral program. Harvard report EP-006 ↗

Evidence of a research proposal is not evidence of an experiment. Harvard's report reproduces Epstein's 2005 Visiting Fellow application describing an interest in group behavior, a proposed timetable for research tests and possible fMRI use. The review says he did little to pursue the course of study and does not establish that he conducted these proposed experiments. Harvard report, PDF p. 6 ↗

Harvard documented inaccurate statements about grant funding. Its investigation concluded that Martin Nowak and PED's chief administrative officer provided false information about the source of matching funds for a John Templeton Foundation grant. The review also identified unresolved questions about what Templeton knew and whether the source was material to the award. That institutional finding warrants an accountability record independently of any unsupported MKULTRA accusation. Harvard report, PDF pp. 16–17 ↗

Harvard's own 2020 review confirmed that after Harvard declined further direct Epstein gifts, he continued introducing donors to university researchers and helped facilitate some gifts. Its reviewers reported that the cooperating donors said they independently chose to give their own money; the review found no evidence contradicting their accounts. The document trail now has specific released-file IDs that can be checked against the Harvard finding. Read Harvard's investigation ↗

EFTA00639492 is indexed as a December 2014 Kahn–Nowak donor-letter exchange, including a proposed $4 million gift letter naming the Program for Evolutionary Dynamics. EFTA02091507 is indexed as an earlier request to prepare a donor letter. EFTA02232201 is indexed as a January 2018 meeting-arrangement exchange. These EFTA records are identified through independent public text indexes; their official scans still require verification. Harvard independently recorded the December 29, 2014 $4 million contribution, so the gift itself is documented even while the draft-letter provenance is being checked.

The distinction between a draft request, a signed pledge, a transfer and a recorded university receipt is necessary before constructing a financial network. No finding in this evidence establishes that these Harvard relationships were part of MKULTRA, MKSEARCH or an intelligence-directed human-testing operation.

First, the Epstein record must be identifiable by its published file identifier and page, with enough surrounding context to distinguish Epstein's own correspondence from a quoted article, an attachment, an allegation, or a third party's speculation. For historic programs, the match must then be checked against a CIA document, a subproject's financial records, or congressional testimony that records the same person, institution, or specific operation. Ambiguous name matches, especially those involving common surnames or altered spelling, remain unresolved until disambiguated.

Second, the date of the proposed activity has to fit the claimed program. MKULTRA's central research activity was mid-twentieth century; the successor MKSEARCH is documented as funded through fiscal 1972. A relationship involving Epstein in the 1990s or 2000s cannot be classified as participation in those historical programs solely through similarity in methods or the existence of former employees. A demonstrable later organizational chain would need independently sourced evidence of who funded, directed or conducted an identifiable operation.

Third, the document's role controls the claim. A witness's allegation is recorded as an allegation, a communications log records communication, and a grant record establishes the specific grant. Only an operational file or an equivalent combination of credible records can establish who actually directed experimentation. Individuals who disclosed abuse or were listed as witnesses are excluded from speculative suspect networks. Redacted personal information will not be reconstructed or republished.

A CIA briefing-book page covering Subproject 90 names MIT's Center for International Studies during 1958–1959. It also contains an explicit clarification about who performed the research and how it was paid for. That correction has to accompany every visual or textual comparison with MIT's later Epstein relationships.

The 1958–1959 CIA summary describes a study of foreign scientists who might encounter American researchers, including the development of approaches for assessing and influencing those scientists. It states that funding of approximately $12,000 passed through the Society for the Investigation of Human Ecology and labels the participant unwitting. Its clarification says the research was carried out by an independent researcher and that the project was neither administered nor funded through MIT. Original CIA briefing scan / PDF page 253 ↗

MIT's separate, much later institutional review concerns Epstein donations between 2002 and 2017. The research described in Subproject 90 preceded Epstein's documented university gifts by more than four decades and involved a different funding channel. No participant, officer, grant, or operational relationship has been shown to connect the two source sets. Their shared institutional name is useful for a disambiguated archival comparison and insufficient for a claim of program continuation. MIT institutional report ↗

The Justice Department's Epstein Library provides DOJ disclosures and access to House material. Its search interface specifically cautions that handwritten or otherwise difficult documents may be missing from electronic results or indexed inaccurately. The House Oversight Committee also recorded a separate September 2025 release of 33,295 Epstein-related pages supplied by DOJ. Each official source permits a traceable chain from a claim to a produced file, with the release date and any redactions preserved.

This draft indexes the official repositories and the two institutional reviews. It does not represent that every released PDF has been read or searched, and no keyword-only result will be promoted into a finding without inspection of the original page. The next documentary acquisition stage is the file-level crosswalk for the names and entities established in the CIA record; claims lacking a matching Epstein document identifier stay in the unresolved queue. DOJ collection EP-001 ↗

On January 30, 2026, the Department of Justice announced more than three million additional responsive pages released under the Epstein Files Transparency Act, bringing the cumulative publication to almost 3.5 million pages. The department stated that the added publication included over 2,000 videos and 180,000 images. Its release statement identifies the case and investigative collections from which records were gathered. Those details supersede any suggestion that the September 2025 House release represents the full public corpus. DOJ January 2026 publication statement ↗

The larger corpus increases the need for file-level identifiers and extraction checks. The DOJ’s own search page warns that some document formats cannot be reliably searched. The present Atlas has identified the release portals and independently reviewed institutional reports; it has not completed a page-by-page audit of this multimillion-page set. No program relationship is inferred from the release volume or from a term appearing in a secondary attachment.

Original CIA correspondence from 1999 to 2011, released in September 2026, shows two personal records requests. In July 2011 the CIA reported finding no open or acknowledged affiliation records within the period it searched. It separately declined to confirm or deny the existence of any classified records. Neither answer establishes a CIA or MKULTRA operational relationship.

The underlying record package includes information unsuitable for display in this Atlas. Only a privacy-screened summary is linked: CIA correspondence analysis. Original reply inspected in the private evidence register as EP-014.