As the United States prepares for the 2024 presidential election, OpenAI has disclosed its strategies to counter misinformation surrounding elections on a global scale. The focus of these efforts is to enhance transparency regarding the origin of information, primarily through the utilization of cryptography aligned with the standards set by the Coalition for Content Provenance and Authenticity.
OpenAI aims to encode the provenance of images generated by DALL-E 3, a move intended to facilitate the identification of AI-generated images using a provenance classifier. This approach will empower the platform to assist voters in evaluating the reliability of specific content related to elections.
The incorporation of cryptography to encode provenance mirrors a similar initiative by DeepMind, known as SynthID, which involves digitally watermarking AI-generated images and audio. This aligns with Google’s broader strategy for election content, unveiled last month. Meta’s AI image generator also employs an invisible watermark in its content, although the company has not yet communicated its preparedness in addressing election-related misinformation.
OpenAI plans to collaborate with journalists, researchers, and platforms to gather feedback on its provenance classifier. In a related move, users of ChatGPT will soon experience real-time global news with proper attribution and links. Additionally, when users inquire about procedural aspects of voting, they will be directed to CanIVote.org, the official online source for U.S. voting information.
Reaffirming its commitment to combating misinformation, OpenAI emphasizes its existing policies to thwart impersonation attempts, such as deepfakes and chatbots. The company explicitly prohibits the development of applications for political campaigning. Users can report potential violations, and OpenAI’s new GPTs offer mechanisms for addressing such concerns.
The success of these initial measures will guide OpenAI in extending similar strategies globally, with the company pledging further announcements in the coming months. The proactive stance underscores OpenAI’s commitment to fostering transparency and countering misinformation in the evolving landscape of digital content.
