Key Points
- A bipartisan group of Senators have proposed a bill to mandate watermarking of AI-generated deepfake content.
- The bill, known as COPIED, aims to enhance transparency, protect creators, and regulate AI content generation.
A proposed legislation, introduced by a cross-party group of Senators, is targeting the misuse of AI deepfakes by requiring such content to be watermarked.
The bill is the brainchild of Senators Maria Cantwell, Marsha Blackburn, and Martin Heinrich, and it suggests a uniform approach to watermarking content produced by AI.
The COPIED Act
The Content Origin Protection and Integrity from Edited and Deepfaked Media Act (COPIED) is designed to strengthen the rights of creators and put restrictions on the kind of content AI can be trained on.
Cantwell believes the bill will provide much-needed transparency into AI-generated content and give creators, such as local journalists, artists, and musicians, more control over their content.
If approved, the bill will also necessitate AI service providers like OpenAI to mandate users to include information about the content’s origin. This information must be machine-readable and cannot be bypassed or removed with AI tools.
The Federal Trade Commission (FTC) will be in charge of enforcing the COPIED Act, treating violations as unfair or deceptive acts, similar to other breaches under the FTC Act.
AI and Ethical Concerns
The introduction of AI has sparked discussions about its ethical implications, given its capability to gather large amounts of data from the internet.
These concerns were highlighted when Microsoft decided against taking up board seats at OpenAI.
Senator Blackburn pointed out that AI has allowed malicious actors to create deepfakes of individuals, including those in the creative community, to mimic their likeness without their consent and profit from counterfeit content.
The bill is being proposed at a time when frauds and scams using Deepfake content have risen by 245%. According to a report from Bitget, these scams are expected to result in losses of $10 billion by 2025.
In the cryptocurrency space, scammers have been using AI to impersonate well-known personalities like Elon Musk and Vitalik Buterin to trick users.
In June 2024, a user of the cryptocurrency exchange OKX lost over $2 million after attackers used deepfake videos of the victim to bypass the platform’s security. The previous month, authorities in Hong Kong shut down a scam platform that used Elon Musk’s likeness to deceive investors.
Recently, Google received criticism from National Cybersecurity Center (NCC) founder Michael Marcotte for its insufficient measures against crypto-targeted deep fakes.
