OpenAI is deploying tools to prevent deepfakes from interfering in US elections, including a ChatGPT feature launched on October 11, 2026 that targets realistic images related to polling locations, ballots, and election-day logistics. The company has formed partnerships with Democracy Works and the Associated Press, and is deploying SynthID watermarking along with C2PA metadata standards. It has also supported the Protect Elections from Deceptive AI Act and disrupted foreign influence operations linked to Iran and Russia. However, testing by the Brennan Center for Justice in September 2026 found that determined users could circumvent existing safeguards. The effectiveness of the new restrictions in addressing the identified gaps remains to be seen ahead of the November 2026 midterms.
Source: Read the original article
