Despite the accord, there is no federal law in the U.S. banning deepfakes. However, 10 states have enacted statutes criminalizing them, and federal agencies are taking enforcement action where possible. The European Union's AI Act requires all AI-generated content to be clearly labeled, and its Digital Services Act is being used to force the tech industry to curb deepfakes. Meanwhile, the number of deepfakes has increased by 900% year over year, and recent polls show that a majority of Americans are concerned about the spread of misleading deepfakes.
Key takeaways:
- Tech companies including Microsoft, Meta, Google, Amazon, Adobe, IBM, OpenAI, Anthropic, Inflection AI, ElevenLabs, Stability AI, X (formerly Twitter), TikTok, Snap, Arm, McAfee, and TrendMicro have pledged to fight election-related deepfakes and have signed an accord to adopt a common framework for responding to AI-generated deepfakes intended to mislead voters.
- The companies will use methods to detect and label misleading political deepfakes, share best practices, and provide swift and proportionate responses when deepfakes start to spread. They will also pay special attention to context in responding to deepfakes.
- No federal law in the U.S. bans deepfakes, but 10 states have enacted statutes criminalizing them. The FTC is seeking to modify an existing rule to cover all consumers, including politicians, and the FCC moved to make AI-voiced robocalls illegal.
- Deepfakes continue to proliferate, with a 900% increase in creation year over year. Recent examples include AI robocalls mimicking U.S. President Joe Biden’s voice and AI-generated audio recordings impersonating a liberal candidate in Slovakia's elections. Surveys show high levels of concern among Americans about the spread of misleading video and audio deepfakes.