Overview

• DeepMind's analysis shows 27% of AI misuse cases aim to influence public opinion through deepfakes.
• Political figures and celebrities are frequent targets of AI-generated fake media.
• Deepfakes are nearly twice as common as other forms of AI misuse, such as cyber attacks.
• Social media platforms struggle to identify and remove deepfake content, raising concerns about voter manipulation.
• The study examined 200 incidents of AI misuse from January 2023 to March 2024.