Disinformation and digital manipulation pose a grave threat to democracy everywhere. Now, the rise of deepfakes– non-consensual AI-generated images, audio and video which can be convincingly realistic and deceptive in nature– are plaguing your platforms, confusing voters in the U.S. and across the world, and undermining people’s trust in factual information. We’re already seeing the way these realistic deepfakes are being weaponized by bad faith actors to spread disinformation, perpetuate sexual abuse, fuel fraud, and erode trust.

We write to you today to specifically call attention to the danger of deepfakes for democracy. From robocalls impersonating President Biden’s voice in New Hampshire to Elon Musk elevating a deepfake video with AI-generated audio of Vice President Kamala Harris, compelling election deepfakes are already reaching and deceiving voters. On top of that, candidates for office are alreadyclaiming that real video or audio of them is AI-generated, leaving voters increasingly uncertain and unable to distinguish truth from fiction.

As civil society organizations and allies dedicated to holding Big Tech accountable, we urge you to take immediate action to implement policies to stop deepfakes from interfering in elections. The harms are not merely imminent; they are already here and felt across the world. Your platforms, with their immense reach and influence, have a responsibility to protect your users from this dangerous form of digital deception.

We ask that you:

• Implement robust detection and moderation systems before the U.S. presidential election specifically designed to identify and prohibit non-consensual and deceptive deepfakes of election officials, election processes, and candidates, in federal, state and local elections, while protecting free speech on platforms in the form of innocuous entertainment or satire easily recognized as manipulated media.
• Require any political AI-generated content to be clearly labeled as such, including a disclosure from the creator about the AI tool used to generate the content for traceability purposes.
• Implement similar systems for other democracies with elections occurring in the later half of this year.
• Collaborate with researchers to provide civil society, academic researchers, and journalists access and insight into the spread of and your enforcement against deceptive electoral deepfakes.

The time for action is now: You must take urgent steps to mitigate against the harm to democracy that deepfakes have the power to yield.

We urge Meta, Snap, Tiktok, YouTube and X to lead the way for all social media platforms to safeguard truth, trust, and accountability and protect our democracy in the digital realm.

Signed,

Accountable Tech
Access Now
Advancing Justice – AAJC
AFT
Center for Countering Digital Hate
Check My Ads
CivAI
Clean Elections Minnesota
Common Cause
Ekō
Electronic Privacy Information Center (EPIC)
Emerge
Free Press
Friends of the Earth Action
Future of Life Institute
GLAAD
Global Witness
Greenpeace USA
Human Rights Campaign
InfoEpi Lab
Interfaith Center on Corporate Responsibility
Investor Alliance for Human Rights
Issue One
Kairos
Media Justice
Media Monitoring Africa
North Carolina For the People
Priorities USA
Public Knowledge
Real Facebook Oversight Board
Reproaction
Secure Elections Network
76 Words
Supermajority
The Sparrow Project
UltraViolet
United We Dream
Unity is Strength
Verified Voting
Vet Voice Foundation