Public records show the agency wanted to build a “meme detector” and use other AI to detect deepfakes.

The recommendation from the U.S. Copyright Office comes as a bipartisan group of senators officially introduced the NO FAKES Act.

The roadmap for artificial intelligence policy encourages the executive branch and appropriators to support $32 billion in annual innovation funding.

The Protecting Consumers from Deceptive AI Act also charges NIST with the development of standards to identify and label AI-generated content.

The legislation would require disclosure of AI-generated content and direct the Federal Election Commission to deal with violations quickly.

Sources tell FedScoop that the Washington Democrat will introduce a series of bipartisan bills related to artificial intelligence issues in the coming weeks.

As Washington putters on AI watermarking legislation, TikTok and Adobe are leading the way with transparency standards.

Rep. Gerry Connolly, D-Va., said additional funding of DARPA and NSF is “critical” to creating advanced and effective deepfake detection tools.

The Protect Elections from Deceptive AI Act offers one path to legislation that would curtail the use of deepfakes that threaten elections.

Existing agreements are often old and fail to account for agencies' need to repurpose and reuse the data.