OpenAI develops a deepfake image detector with high accuracy, collaborating with industry for transparency

From SiliconANGLE: 2024-05-07 21:12:21

OpenAI has developed a detector to identify deepfake images created by its DALL-E AI generator, aiming to combat the spread of manipulated content. The detector boasts over 98% accuracy with DALL-E images but struggles with images from other generators, hinting at the challenges ahead. Progress will be slow, but industry collaborations, like the C2PA standard with Google, aim to enhance content provenance and transparency online, driven by fears of AI manipulation in politics and legislation. Engagement and transparency are key in combating the deepfake crisis, as OpenAI and other industry players work towards safeguarding digital content integrity.



Read more at SiliconANGLE: OpenAI announces its new tool that can detect deepfake images