The FBI is exploring the use of artificial intelligence to verify signatures on mail-in ballot envelopes seized from Fulton County, Georgia. This initiative is part of an ongoing investigation into the 2020 election, which has been characterized by claims of voter fraud, particularly in Georgia where President Donald Trump lost by 11,779 votes. The FBI collected approximately 700 boxes of election materials, including around 150,000 mail-in ballots, of which about 116,000 were cast for Biden.
Internal communications indicate that the FBI's efforts involve comparing signatures on ballot envelopes with those on election documents like registration forms. The status of this signature-matching initiative is currently unclear, and the FBI has redirected 260 analysts to focus on the Fulton County investigation.
Concerns have been raised regarding the accuracy of signature matching, with experts noting that it can lead to high rates of rejected ballots, particularly affecting voters of color and other marginalized groups. A political scientist previously testified that for every illegitimate ballot identified, 32 legitimate ballots were blocked. The FBI's technology specialist acknowledged that while AI could provide some level of accuracy, the effectiveness of the analysis would depend on the guidelines established for identifying fraud.
The discussion around signature matching has gained attention since the 2020 election, with various states implementing stricter laws regarding signature verification. Critics argue that these measures can disproportionately affect certain voter demographics and may not effectively prevent fraud. The FBI's ongoing investigation and the use of AI in this context continue to draw scrutiny regarding potential political influences on the findings.