OpenAI blocks over 250,000 image requests tied to 2024 presidential candidates
- OpenAI blocked over 250,000 image generation requests for U.S. presidential candidates.
- Concerns over misinformation from generative AI intensified with a reported 900% rise in deepfakes.
- With ongoing challenges, experts warn voters against sourcing election information from AI due to accuracy issues.
In the lead-up to the 2024 U.S. presidential elections, OpenAI reported that it had rejected over 250,000 requests to generate images of various candidates. The request rejections included those related to political figures such as President-elect Donald Trump, Vice President Kamala Harris, President Joe Biden, Minnesota Governor Tim Walz, and Vice President-elect J.D. Vance. This significant number of rejections underscores the company's commitment to proactively manage the potential risks of misinformation generated by its AI technologies. The increase in generative AI usage has raised alarms regarding the potential for deepfakes and misinformation to influence elections. According to data from Clarity, the incidence of deepfakes surged by 900% in the past year, with some instances linked to efforts by foreign actors aiming to interfere with U.S. democracy. U.S. intelligence has pointed out that Russian entities have allegedly funded misinformation campaigns, which further heightens the urgency to combat such threats. OpenAI's actions come as part of a broader effort to regulate AI systems in a landscape increasingly filled with misinformation. In an extensive October report, the company emphasized that it had interrupted more than 20 deceptive operations that attempted to utilize its AI models for spreading false narratives. However, the report noted a silver lining: none of the election-related misinformation campaigns conducted using OpenAI's products received significant viral attention or lasting engagement, illustrating some effectiveness of the barriers implemented. As concerns about misinformation grow, lawmakers and experts have cautioned voters against relying on AI-powered chatbots for accurate electoral information. Alexandra Reeve Givens, CEO of the Center for Democracy & Technology, advised that voters should be wary due to the current unreliability and inaccuracy associated with these large language models. The situation highlights the need for careful monitoring of AI technologies in the political domain, especially in the context of ongoing elections both in the U.S. and around the world.