ChatGPT rejected over 250,000 AI image requests in the month leading up to Election Day according to a Friday blog post from OpenAI.
Users tried to generate photos of President-elect Trump, Vice President Harris, Vice President-elect Vance, President Biden and Gov. Tim Walz (D-Minn.)
“We’ve applied safety measures to ChatGPT to refuse requests to generate images of real people, including politicians,” the company wrote. “These guardrails are especially important in an elections context and are a key part of our broader efforts to prevent our tools being used for deceptive or harmful purposes.”
Meta rolls out new feature in hopes of protecting users
Instead, ChaptGPT directed election related questions to CanIVote.org through a partnership with the National Association of Secretaries of State (NASS) in an effort to remain politically neutral.
The platform is directing inquiries about the election results to news outlets like the Associated Press and Reuters. In recent weeks, OpenAI had to ban a foreign influence operation known as Storm-2035 which was pushing out influential Iranian political content.
The company said it will continue to monitor ChatGPT to ensure responses are accurate and ethical. This year, the company applauded the Biden administration’s efforts to outline policy regarding national security and AI technology.