Meta Reports Minimal Generative AI Impact on 2024 Elections
Meta's year-end report reveals that generative AI had minimal impact on the 2024 elections across its platforms, including Facebook, Instagram, and Threads. This analysis encompasses major elections worldwide, from the U.S. and India to Brazil and the EU Parliament. The company's existing policies and processes proved effective in mitigating risks associated with AI-generated content.
Fact-checked misinformation related to elections, politics, and social issues constituted less than 1% of all content reviewed. Meta's Imagine AI image generator played a crucial role by rejecting over half a million image creation requests featuring prominent political figures, effectively preventing the spread of deepfakes. Cyber Monday sales trends show a similar focus on preventing misuse of technology.
Furthermore, coordinated disinformation campaigns saw limited success in leveraging AI for content generation. Meta's strategy of targeting account behavior rather than individual posts, regardless of AI involvement, proved effective in disrupting these campaigns. Around 20 covert influence operations were shut down globally, primarily due to their lack of authentic engagement and use of fake likes and followers. This echoes concerns about data brokers and location data abuse.
While Meta highlights its success in managing AI-related election interference, it points to other platforms like X and Telegram as sources of false election-related videos linked to Russian influence operations. The company plans to review its policies and announce any updates in the coming months. For more on tech policy, see CFPB's plans to regulate data brokers.