{"resourceId":"openai-meta-federal-evidence-committee-deepfake-warning-20261015","versions":[{"version":"external-b94a0f6529916f1322fdde17e640df47a1513b240b2d653702fe72ec0cd8e333","resource":{"id":"openai-meta-federal-evidence-committee-deepfake-warning-20261015","title":"OpenAI and Meta tell the federal Advisory Committee on Evidence Rules that deepfakes are becoming undetectable, as the panel weighs draft deepfake and AI-evidence rules; only 15 of 931 surveyed judges had seen a deepfake challenge","organization":"U.S. Judicial Conference Advisory Committee on Evidence Rules","sector":"Courts / evidence authentication and AI-generated evidence","geography":"United States (federal courts; influences state evidence rules)","publishedAt":"Thu, October 15, 2026 at 11:44 a.m. PDT (Investing.com via Yahoo Finance)","publicationDate":"2026-10-15","eventDate":"2026-10-15","sourceName":"Investing.com (Yahoo Finance Canada)","sourceLabel":"OpenAI and Meta warn U.S. judicial panel about deepfake evidence","sourceUrl":"https://ca.finance.yahoo.com/news/openai-meta-warn-u-judicial-184412718.html","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["governance-procurement","data-security","knowledge-work"],"finding":"At the committee’s Oct. 15 meeting in Boston, OpenAI policy counsel Shira Anderson said she does not believe a scientific method to detect AI images will exist, and that because each company uses its own watermarking system, no universal method can verify all content. Meta’s Devin Norder said recent AI media cannot be distinguished by humans and is cheap and fast to create. The committee has worked since 2023 on deepfake authentication and on holding AI-generated evidence to expert-witness reliability standards. A Federal Judicial Center survey in March found only 15 of 931 responding judges had encountered a deepfake challenge, and chair Judge Jesse Furman noted the issue has not yet appeared often in cases.","sledRelevance":"Body-camera, surveillance, jail-call and citizen video are core evidence in criminal courts. Vendor testimony that detection and watermark checks will not be reliable shifts the burden to provenance and chain-of-custody controls at the agencies that collect evidence.","evidence":"Opened the Investing.com report syndicated on Yahoo Finance. Quotes and the FJC survey figure are as reported. Secondary analysis (On The Wire, Oct. 19) reviewed during research describes draft Rule 901(c) and revised Rule 707 as not adopted; those rule details are not drawn from the opened source.","architectureImplications":"Interpretation: agencies should rely on capture-time provenance (signed device metadata, hash on ingest, immutable evidence logs) rather than after-the-fact deepfake detectors.","governanceImplications":"Federal rule changes remain under study; the committee has not adopted a deepfake rule. State courts often follow federal evidence rules, so the outcome will shape state practice.","securityPrivacyImplications":"Vendor watermarking is fragmented, so authenticity checks cannot depend on one company’s system.","caveats":"Brief wire report; statements are company representatives’ views; survey shows challenges are still rare; no rule adopted.","streamIds":["public-safety"],"roles":{"sales":"Sales framing: the companies building generative models told federal rule-makers that deepfake detection will not be reliable. For prosecutors and police evidence units, the bounded offer is an evidence-provenance review, not a detection product. The source shows challenges are still rare (15 of 931 judges), so urgency claims must be modest and no tool can be sold as a guaranteed authenticity check.","engineering":"Engineering framing: if detection is unreliable, authenticity depends on provenance captured at the source. Pre-sales should check whether cameras and evidence platforms hash files on ingest, preserve original metadata, log every access and export, and can produce a chain-of-custody report a court will accept. Proof should test that report on real closed cases.","delivery":"Delivery framing: the work is documenting and strengthening evidence provenance across body-camera, surveillance and citizen-submitted media. Dependencies include evidence platform features and prosecutor input on what courts require. The owner is the evidence unit with the prosecutor’s office. Acceptance should track complete chain-of-custody records rather than detector scores."},"roleDetails":{"sales":{"customerProblem":"Courts may soon scrutinize digital evidence authenticity while vendors say deepfake detection will not be reliable.","stakeholders":["District attorney","Police evidence unit manager","Public defender","Court administrator","Agency CIO"],"discoveryQuestions":["How do you prove a video has not been altered?","Do you accept citizen-submitted media, and how is it verified?","Are files hashed on ingest?","Have any authenticity challenges occurred?"],"valueHypothesis":"Stronger provenance records can prepare agencies for authenticity challenges without relying on detectors.","potentialEngagement":"4-week evidence provenance and chain-of-custody gap assessment for digital media.","unsupportedClaims":["A detector can reliably identify deepfakes","Deepfake challenges are common today","A specific federal rule will be adopted"]},"engineering":{"fit":"Prosecutors and agencies managing large volumes of digital media evidence.","architectureIntegration":"Hash on capture or ingest, preserve original files and metadata, log access and exports, and generate chain-of-custody reports from the evidence platform.","prerequisites":["Evidence platform with hashing and audit logs","Policy for citizen-submitted media","Prosecutor requirements"],"constraints":["Rules still under study","Mixed device fleets lack signed metadata","Citizen media lacks provenance"],"security":"Restrict edit rights, keep originals immutable and log every transformation such as redaction.","proofOfValue":["Proposed: generate chain-of-custody reports for five closed cases","Proposed: verify hashes match originals after redaction workflows"]},"delivery":{"work":"Inventory evidence sources, close provenance gaps and create a standard authenticity packet.","dependencies":["Evidence platform configuration","Prosecutor guidance","Device metadata support"],"ownership":"Evidence unit manager with the prosecutor’s office","skillsAdoption":"Train evidence technicians and attorneys on provenance documentation and authenticity challenges.","governanceCheckpoints":["Prosecutor review of packet format","Annual evidence audit","Review after federal rule decisions"],"acceptanceCriteria":["Proposed: 100% of digital media items hashed on ingest","Proposed: authenticity packet produced within one business day on request"],"risks":["Citizen media without provenance","Legacy devices lacking metadata","Overreliance on detector vendors"]}},"retrievedAt":"2026-10-20T02:53:19Z","enrichedAt":"2026-10-20T02:58:42Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Interpretation: prosecutors, defenders and evidence technicians will need training on authentication challenges and provenance documentation.","procurementImplications":"Interpretation: evidence-management and camera contracts should require tamper-evident hashing, signed metadata and exportable chain-of-custody logs.","operatingModelImplications":"Evidence units and prosecutors share responsibility for proving authenticity; courts set the threshold.","sourceVerification":{"openedUrl":"https://ca.finance.yahoo.com/news/openai-meta-warn-u-judicial-184412718.html","referenceExcerpt":"A survey conducted by the Federal Judiciary Center in March found that only 15 out of 931 judges who responded had encountered a case","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}