Ahead of the 2024 New Hampshire presidential primary, thousands of voters picked up the phone and heard what sounded like President Joe Biden telling them to skip the primary and “save their vote for the November election.”
It was not Biden. It was an AI-generated clone of his voice. A political operative named Steven Kramer admitted he had arranged the calls.
In June 2025, a New Hampshire jury acquitted him on all 11 felony counts of voter suppression and all 11 counts of impersonating a candidate. Kramer testified that he made the calls “as a warning about the dangers of artificial intelligence.” Separately, the Federal Communications Commission had fined him $6 million in September 2024, and the telecom company that carried the calls paid $1 million. New Hampshire Public Radio, June 13, 2025.
A man admitted faking the president’s voice to tell citizens not to vote, and a jury found him not guilty. Whatever you think of that verdict, it tells you how unprepared our institutions are for what is coming.
The details make it harder, not easier, to shrug off. According to NHPR, Kramer had previously worked for the presidential campaign of Rep. Dean Phillips, whose team denied any involvement in the calls. New Hampshire’s attorney general, John Formella, responded to the acquittal by saying his office would “continue to work diligently to address the challenges posed by emerging technologies, including artificial intelligence, to protect the integrity of our elections.” That is a reasonable promise. It is also an admission that the current tools did not produce a criminal conviction in a case where the man behind the calls never denied making them. Every operative watching learned the same lesson.
The public is worried, and it is right to be
The Pew Research Center surveyed Americans and AI experts in 2024. It found that 66% of U.S. adults were highly concerned about people getting inaccurate information from AI, a worry shared by 70% of experts. A majority of the public was highly concerned about AI being used to impersonate people. And 62% of Americans said they lacked confidence that the government would regulate AI effectively. Pew Research Center, April 3, 2025.
The gap between the builders and the public is striking. Forty-seven percent of AI experts said they were more excited than concerned about AI’s growing role in daily life. Only 11% of the public agreed. Fifty-one percent of Americans said they were more concerned than excited.
The crime data backs up the public’s instinct. The FBI’s report on 2025 counted 22,364 complaints tied to AI, with nearly $893 million in losses. The bureau said scammers were using “fake social profiles, voice clones, identification documents, and believable videos depicting public figures or loved ones.” FBI, April 6, 2026.
Two dangers, not one
Most people think the danger of deepfakes is that we will believe something fake. That is the first danger, and it is real. A fake voice can steer an election. A fake video can ruin a reputation before anyone checks it.
The second danger is quieter and may be worse: we will stop believing things that are true.
Once everyone knows convincing fakes exist, anyone caught on a real recording can simply claim it was AI. A politician on tape taking a bribe. An executive caught lying. A public figure saying something ugly. “That’s a deepfake” becomes an all-purpose escape hatch. Every piece of genuine evidence gets dragged down to the level of the fakes.
That is the real threat to a society’s moral fabric. Accountability depends on evidence. Courts, journalists and voters all rely on the idea that a recording, a photo or a video can show what actually happened. If people conclude that nothing can be trusted, the powerful gain the most, because they no longer have to answer for what the camera caught.
Where the law is catching up
The legal system is moving, slowly.
- Robocalls: In early 2024, the FCC unanimously ruled that calls using AI-generated voices count as “artificial” under the Telephone Consumer Protection Act, giving state attorneys general new tools to pursue them. FCC.
- Intimate deepfakes: The federal TAKE IT DOWN Act now requires covered platforms to remove nonconsensual intimate images, including AI-generated deepfakes, within 48 hours of a valid request. The FTC began enforcing it on May 19, 2026. People can report platforms that fail to comply at TakeItDown.ftc.gov. FTC Consumer Alert, May 2026.
Those are real steps. But the Kramer case shows the limits. Regulators fined him millions. A jury still acquitted him of the crimes charged. The law has not settled how to treat someone who uses a synthetic voice to mislead voters while claiming it was a public service announcement.
How to verify without giving up on evidence
The answer to deepfakes is not to believe nothing. It is to verify better. Here is a working standard for any shocking clip that crosses your feed:
- Find the original. Where did the video or audio first appear? A clip with no source, posted by an anonymous account, deserves suspicion.
- Look for corroboration. Did credible outlets or officials independently confirm it? Did anyone who was present say it happened?
- Check the context. Real clips are often edited to change their meaning. Look for the full recording.
- Be most skeptical when it confirms what you already believe. That is exactly when you are easiest to fool.
- Do not share until you know. Every repost of a fake does the forger’s work for free.
And apply the same standard in reverse. When someone you support is caught on a recording and immediately claims “it’s AI,” ask the same questions. Is there evidence it was altered? Or is the fake-news defense just a way to escape accountability?
The verdict
Artificial intelligence did not invent lying. It made lying cheaper, faster and more convincing, and it handed every guilty person a new excuse. The Moral Decay Index tracks trust in institutions, and trust in evidence is the foundation under all of them.
The public is right to be worried. But worry is not a plan. The plan is simple, even if it takes effort: verify before you believe, verify before you dismiss, and hold your own side to the same standard of proof you demand from everyone else.
Sources
- New Hampshire Public Radio, “Political operative who admitted to creating fake Biden robocalls found not guilty,” June 13, 2025
- Pew Research Center, “How the U.S. Public and AI Experts View Artificial Intelligence,” April 3, 2025
- Federal Bureau of Investigation, “Cryptocurrency and AI Scams Bilk Americans of Billions,” April 6, 2026
- Federal Communications Commission, “FCC Makes AI-Generated Voices in Robocalls Illegal”
- Federal Trade Commission, “What will the FTC’s enforcement of the TAKE IT DOWN Act mean for you?” May 2026
If this made you angry, good. Now turn it inward. The Moral Consistency Journal is a 30-day audit of the places where your own standards bend. More Case Files and data breakdowns live at themoraldecayindex.com. Books that go deeper are in our Bookshop.org store, which supports independent bookstores.