FBI Explored Using AI to Analyze Signatures on 150000 Seized Fulton County Mail-In Ballots
The FBI has explored using artificial intelligence to compare signatures on roughly 150,000 mail-in ballot envelopes seized from Fulton County, Georgia against signatures on registration forms and other election records, according to internal communications and an agency tech specialist cited by ProPublica, as part of the Trump administration's reinvestigation of the 2020 election, in which Fulton County went for Biden by roughly 116,000 of the seized ballots
The FBI's consideration of commercial AI models to adjudicate signature validity on a politically charged ballot cache outsources forensic judgment on a method forensic scientists agree lacks scientific consensus at scale. Internal dissent among bureau staff over methodology, set against continued pressure from leadership to pursue findings, indicates the outcome will be shaped as much by political expectation as technical validity, a dynamic that recurs alongside Patel's direct role in the White House leak probe and points to a director engaged across politically directed investigations bureau-wide. ProPublica's reporting rests on internal communications and a single unnamed technology specialist, with no independent corroboration; the discussions could equally reflect routine technology due-diligence rather than a committed operational plan, and the bureau's own sourcing leaves the initiative's status unresolved since late June.
3 sources
- FBI Has Looked at Using Questionable AI Tech to Review Signatures on Seized Mail-In Ballots -
ProPublica - FBI Has Looked at Using Questionable AI Tech to Review Signatures on Seized Mail-In Ballots -
Election Law Blog - FBI Explores Using AI To Check Validity Of Signatures On Thousands Of Seized 2020 Georgia Mail-In Ballots -
Joe.My.God.