Stanford Internet Observatory and similar research groups have documented multiple influence operations in the past year that deployed AI-generated text, images, and social media accounts to manipulate political opinion on elections, conflicts, and health policy in the United States and allied nations. The operations are significantly more sophisticated than earlier influence campaigns and produce content that is much harder to identify as inauthentic.
The FBI's foreign influence task force has expanded significantly in response to the threat, but the volume of potential AI-generated influence content is overwhelming traditional monitoring capabilities. Researchers are developing AI detection tools specifically for influence operation content, but the detection arms race consistently favors offense. Platform companies are implementing new restrictions on AI-generated political advertising, but organic content manipulation remains largely unchecked.
