Three deep-fake detection software experts testified to the Senate subcommittee on privacy, technology and legislation on Tuesday about how Congress can regulate the artificial intelligence space, particularly the rise of harmful deep-fakes.
Deepfakes are misleading audio, video or images created or edited with artificial intelligence technology.
in January, deepfake audio A robocall impersonating President Joe Biden has been sent to thousands of voters telling Democrats not to vote in the New Hampshire primary. Other profound frauds Use AI to “undress”. real photos of women and girls juxtaposing real faces with fake nudity. Both types of deep fraud were discussed during Tuesday’s hearing, which focused on disinformation during the 2024 election cycle.
The committee’s chairman, Sen. Richard Blumenthal, D-Conn., said the regulatory framework proposed by Congress would require independent testing of AI technology before it is released to the public, along with potential penalties.
“We must make no mistake: The threat of deep political fraud is real, it is happening now,” he said. “The future, possibly or hypothetically, is not science fiction. “Artificial intelligence is already being used to interfere in our elections, spread lies about candidates and prevent voting.”
Zohaib Ahmed, CEO and Co-Founder of Resemble AI; Ben Colman, CEO and co-founder of Reality Defender; and DeepMedia CEO Rijul Gupta testified in court.
Sen. Amy Klobuchar, D-Minn., said 11 states have passed laws banning deep-rigged election interference. Other states have enacted laws to create civil remedies for victims of deep-fake sexual abuse. But proposed federal legislation around both issues has stalled home and Senate Sen. Josh Hawley, R-Mo., urged both party leaders to plan.
“The danger of this technology without guardrails and safety features becomes painfully clear,” Hawley said. “Let’s not allow the companies that control social media technology in this country, that control the news in this country, to use artificial intelligence to further their position in this country and in the political process.”
In the case of the New Hampshire Biden robocall, This was reported by NBC News said the voice-cloned audio was created by a street magician with ties to a rival Democratic campaign. The audio was created with software from ElevenLabs, which anyone can download and use.
NBC News also informed The audio generated with ElevenLabs contains a hidden watermark—several major tech companies have proposed watermarking technology as a solution to deep spoofing—but in the process of creating the robocall, the watermark was removed from the audio received by the primary voters.
“Anyone with a Google search and an internet connection can do just about anything fun and dangerous they can imagine,” says Reality Defender’s Colman. “We’ve seen time and time again that they won’t follow the rules.”