Safeguarding Democracy: The Latest in AI Deepfake Legislation

By: Talia Cole

Understanding the Threat Landscape

From hyper-realistic audio and video fakes to manipulative political content, AI deepfakes pose unprecedented challenges:

• In 2023, deepfake videos surged to nearly 96,000 online—98% of which were pornographic, with 99% targeting women and girls.
• The infamous Taylor Swift deepfake controversy, which drew tens of millions of views, underscored the urgent need for robust regulation.
• In elections, AI-generated media has already interfered with campaigns—like a cloned voice of a Chicago mayoral candidate—highlighting the risks to democratic integrity.

Landmark Legislation: The TAKE IT DOWN Act

Lawmakers responded decisively:

• The TAKE IT DOWN Act (Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks) became law on May 19, 2025.
• It mandates platforms to remove non-consensual intimate imagery (NCII), including AI-generated deepfakes, within 48 hours of a takedown request, and criminalizes their distribution.
• The law received overwhelming bipartisan support: it passed the Senate unanimously, the House with a 409–2 vote, and was signed into law by President Trump.
• First Lady Melania Trump also played a visible role, advocating for protections for young women online.

Why this matters: It marks the first substantive U.S. law targeting AI-generated content, setting a precedent for digital rights and platform accountability.

What’s Next: The No Fakes Act and Broader Oversight

While TAKE IT DOWN addresses NCII, lawmakers and industry groups recognize the need for broader guardrails:

• The proposed No Fakes Act would protect individuals—including public figures—from unauthorized AI-generated likenesses and voice clones.
• Backed by artists, YouTube, and the RIAA, it’s designed to safeguard creative integrity while balancing free speech rights.

State-Level Regulation: Innovation vs. Uniformity

States are also taking action—sometimes in alignment, other times in tension with federal efforts:

• Tennessee’s ELVIS Act protects musicians from AI-generated impersonations.
• New York’s Hinchey Law allows victims to sue creators of explicit deepfakes.
• Texas Senate Bill 20 criminalizes AI-generated child pornography, even in software-generated form.

But controversy exists: in June 2025, a proposed ten-year federal moratorium on state AI laws sparked bipartisan backlash, with critics arguing it would hinder local protections. The Senate struck it down.

Why These Developments Matter

1. Vulnerable populations under threat: Women and girls remain disproportionately targeted.
2. Institutions now have enforcement tools: Schools and platforms are empowered to act swiftly.
3. Precedent for broader regulation: TAKE IT DOWN opens the door for future legislation.
4. Transparency and platform accountability: Platforms are being required to add labels and metadata.
5. Balancing rights and protections: Critics warn against government misuse or censorship overreach.

Final Thoughts: A Policy Landscape in Motion

The U.S. is building legal defenses against AI misuse—especially where democracy and dignity are at stake. But deepfake technology evolves quickly, often faster than laws can keep up. Future policy will need to stay nimble, ensuring both protection and innovation.

---

Works Cited

“99% of Nonconsensual Sexual Deepfakes Target Women and Girls. It’s Time for Congress to Act.” San Francisco Chronicle, 14 Jan. 2025, http://www.sfchronicle.com/opinion/openforum/article/99-nonconsensual-sexual-deepfakes-target-women-21021885.php.

“Artificial Intelligence and Elections.” Wikipedia, Wikimedia Foundation, 2025, en.wikipedia.org/wiki/Artificial_intelligence_and_elections.

“Deepfakes in American Elections.” American Bar Association, 2024, http://www.americanbar.org/groups/public_interest/election_law/american-democracy/our-work/deepfakes-american-elections/.

“ELVIS Act.” Wikipedia, Wikimedia Foundation, 2024, en.wikipedia.org/wiki/ELVIS_Act.

“Regulation of AI in the United States.” Wikipedia, Wikimedia Foundation, 2025, en.wikipedia.org/wiki/Regulation_of_AI_in_the_United_States.

“TAKE IT DOWN Act.” Wikipedia, Wikimedia Foundation, 2025, en.wikipedia.org/wiki/TAKE_IT_DOWN_Act.

“Taylor Swift Deepfake Pornography Controversy.” Wikipedia, Wikimedia Foundation, 2024, en.wikipedia.org/wiki/Taylor_Swift_deepfake_pornography_controversy.

“The State of Deepfake Regulations in 2025: What Businesses Need to Know.” Reality Defender, 19 Mar. 2025, http://www.realitydefender.com/insights/the-state-of-deepfake-regulations-in-2025-what-businesses-need-to-know.

“Texas Senate Bill 20.” Wikipedia, Wikimedia Foundation, 2025, en.wikipedia.org/wiki/Texas_Senate_Bill_20.

“Trump Signs AI Deepfake Act into Law and House Republicans Push Moratorium on State AI Regulations.” ML Strategies, 22 May 2025, http://www.mlstrategies.com/insights-center/viewpoints/54031/2025-05-22-president-trump-signs-ai-deepfake-act-law-and-house.

“U.S. House Approves AI Deepfake Bill, Trump Signs Executive Order on AI.” Mintz, 2 May 2025, http://www.mintz.com/insights-center/viewpoints/54731/2025-05-02-congress-passes-ai-deepfake-law-trump-signs-eo-ai.

“Washington Grapples with State AI Laws Moratorium.” The Washington Post, 22 May 2025, http://www.washingtonpost.com/politics/2025/05/22/state-ai-laws-moratorium-pushback-blackburn.

“What Is ‘Revenge Porn’? Why Melania Trump Is Speaking Out.” Time Magazine, 20 May 2025, time.com/7264316/what-is-revenge-porn-melania-trump/.