By Roman Aglaganov
In January 2024, more than 20,000 New Hampshire voters picked up the phone and heard what sounded like President Joe Biden urging them not to vote in the state’s primary. It was not him. It was an AI-cloned voice, and it became the first big test of what happens when deepfakes enter an election.
Who made the call
Investigators traced the calls to a Texas company and the phone carrier that transmitted them, and then to Steve Kramer, a political consultant. What alarmed everyone was how easy it had become: cloning a famous voice no longer needs a studio, just some recordings and off-the-shelf software.
The regulators move
On 8 February 2024, less than three weeks after the calls, the US Federal Communications Commission voted unanimously that AI-generated voices count as “artificial” under the Telephone Consumer Protection Act, a law from 1991. In plain English: robocalls using cloned voices are illegal without the recipient’s prior consent.
In September 2024 the FCC fined Kramer $6 million.
The twist
Kramer was also charged under New Hampshire law with voter suppression and impersonating a candidate. In June 2025, a jury acquitted him on all counts. The FCC’s $6 million fine still stands, though, so the case ended up as a lesson in which rules bite: old consumer-protection law, applied quickly to new technology, proved more useful than election law.
What to do: if a familiar voice calls with urgent instructions, whether it is a politician or a relative asking for money, hang up and check through another channel. Voices are now easy to fake.
