India and US States Escalate Deepfake Regulation
India’s new IT Rules mandate a three-hour takedown window for harmful AI-generated content, sparking fears of censorship. In the U.S., Massachusetts is considering mandatory disclosure for AI in political advertising after a deepfake audio ad impersonated the governor. The parallel moves highlight a global trend toward stricter regulation of AI-driven election disinformation.
- India's amended IT Rules, effective February 20, 2026, require social media platforms to label all "synthetically generated information" and embed permanent metadata to trace its origin. For certain violations, including deepfakes, the takedown window has been drastically cut from 36 hours to just three. - In the U.S., the Federal Communications Commission (FCC) fined a political consultant $6 million for a robocall that used an AI-generated voice clone of President Joe Biden to discourage voting in the New Hampshire primary. This incident has spurred legislative action in New Hampshire and other states. - The Massachusetts House passed two bills in February 2026 to regulate AI in elections. One bill mandates disclosure on any political ad using synthetic media, carrying a $1,000 fine for violations, while the other prohibits the distribution of "materially deceptive" media within 90 days of an election. - A proposed federal law, the NO FAKES Act, would establish a national framework for protecting an individual's voice and likeness as a property right, allowing legal action against unauthorized digital replicas. - The European Union's AI Act requires the labeling of AI-generated content, with mandatory compliance expected by August 2026. This is six months after India's more aggressive rules take effect. The EU's Digital Services Act also holds large platforms accountable for mitigating the risks of manipulated media. - Penalties for non-compliance in India can include fines and losing "safe harbor" protections, which shield platforms from liability for user-generated content. The rules also link the creation of certain deepfakes to criminal laws like the Bharatiya Nyaya Sanhita and the POCSO Act. - Globally, the approach to deepfake regulation varies. China was a first mover with its "Deep Synthesis Regulations" in 2023, while South Korea criminalized non-consensual deepfake pornography in 2024. In the U.S., more than 45 states have some form of deepfake legislation, often focused on elections or non-consensual intimate imagery.
