Updated
Updated · The New York Times · Nov 3
Researchers Find 19,300 Deceptive Bot Accounts in 6 Midterm Races
Updated
Updated · The New York Times · Nov 3

Researchers Find 19,300 Deceptive Bot Accounts in 6 Midterm Races

3 articles · Updated · The New York Times · Nov 3

Summary

  • Cyabra said coordinated inauthentic accounts targeted six major midterm contests from Aug. 31 to Oct. 19, spanning races in Georgia, Kansas, Maine, Michigan and Ohio.
  • Out of 274,000 posts reviewed across Facebook, Instagram, TikTok and X, roughly 1 in 9 posts came from inauthentic accounts, and bots produced 1 in 4 replies to Georgia governor candidates on X.
  • The activity hit both parties in several races, including the Georgia and Ohio governor contests and the Kansas and Maine Senate races, suggesting an effort to distort discussion rather than boost one side alone.
  • Researchers said some accounts behaved like managed personas rather than simple automation, posting in synchronized bursts and masking their identity, though real users still drew 99.8% of X views.
  • The origin and electoral impact remain unclear, but researchers warned cheaper AI tools could make such influence campaigns a lasting feature of U.S. elections.

Insights

As AI safeguards prove porous, how might undetected deepfakes shape public trust before final election results are certified?
With millions voting early, could shifting weather patterns still silently disenfranchise key demographics on Election Day?