OpenAI is teaming up with major social media and technology companies to tackle one of the thorniest problems of the AI era: keeping elections clean when the tools to fake reality have never been cheaper or more accessible.
The effort centers on the Tech Accord to Combat Deceptive Use of AI in 2024 Elections, signed on February 16, 2024, at the Munich Security Conference. More than 20 companies put their names on the dotted line, including Meta, Google, Microsoft, and TikTok.
The accord commits signatories to improving threat detection tools, enhancing transparency through provenance signals and watermarking, sharing intelligence on emerging threats, and investing in public education about AI-generated deceptive content.
OpenAI has also updated its own policies. Since January 2024, the company prohibits using its tools for candidate impersonation and spreading false voting information. Additionally, OpenAI endorsed the bipartisan Protect Elections from Deceptive AI Act, aimed at setting legal boundaries around AI-generated media in federal political ads.
Compliance reviews scheduled for early 2025 will determine if the voluntary model works or if lawmakers push for binding legislation.