One thing we don’t need after the Capitol insurrection of January 6, 2021 is another danger to democracy. But unlike the televised violence that day, Blumenthal’s hearing showed how artificial intelligence can be used to overturn elections far more covertly than MAGA rioters supporting President Donald Trump had attempted. three years ago.
This is a bipartisan threat that has generated bipartisan resolve.
Two Democrats, Sens. Amy Klobuchar (Minn.) and Chris Coons (Del.), and two Republicans, Josh Hawley (Mo.) and Susan Collins (Maine), are pushing legislation that would ban misleading AI elements in ads policies. . The bill, introduced in September, would also allow candidates for federal office to ask U.S. courts to order the removal of false information and award compensation to the candidates.
But the legislation is not moving quickly and the urgency is clear, as the hearing highlighted.
“Are we going to have to have an election disaster before Congress realizes, ‘Well, we should really do something to give the public some sense of security, some sense of certainty that what they’re seeing and hear is actually real or is it real?’ actually fabricated,” asked Hawley, the top Republican on the subcommittee on Privacy, Technology and the Law, which held last week’s session.
Deepfakes featuring Trump and President Biden have already been used to mislead the public.
Hearing witness David Scanlan, New Hampshire’s secretary of state, recalled that he had “prepared to run a very good” presidential primary there the weekend before the January vote. Then things changed. He began hearing about “an automated call using AI with President Biden’s voice, asking individuals not to vote in the election.” The call appears to come from a phone number associated with a former Democratic Party official.
“It’s important that you save your vote for the November election,” the voice said. “Your vote makes a difference in November, not this Tuesday.”
The message was false, as was the association with the party official.
“This is what suppressing voter turnout looks like,” Blumenthal said after playing the audio during the hearing. The Associated Press said it was “perhaps the first known attempt to use artificial intelligence to interfere with a U.S. election.”
Last month, the BBC reported on fake photos of Trump surrounded by African Americans, apparently broadcast to give a false impression about his level of support for black people. Last year, fake footage linked to Trump’s court appearances showed him brawling with police and wearing prison uniforms.
What’s also worrying is how little effort it takes to deceive people with today’s technology, which makes really good fakes simpler. With free online programs, Blumenthal said, “voice cloning, deepfake images and videos are disturbingly easy for anyone to create.”
The one with Biden was done “by a street magician whose previous claim to fame was holding world records for bending spoons and escaping straightjackets,” Blumenthal added. “And if a street magician can cause this much trouble, imagine what Vladimir Putin or China can do. In fact, they do.
Five years ago, the Washington Post reported on a slick Russian campaign that used social media to turn off black voters, according to documents released by the Senate Intelligence Committee. One poster showed the face of a black man next to the words “I will not vote.” It was primitive compared to today’s efforts.
While deepfakes involving Biden or Trump will receive publicity if discovered, Blumenthal said “local elections pose an even greater risk” due to the worrying decline in local journalism, an issue that the senator explored during a hearing in January.
“When a local newspaper is closed or understaffed, there may be no one to check the facts, no one to publish those Pinocchio pictures and no one to correct the facts,” he said last week. “This is a recipe for toxic and destructive politics.”
Additionally, a report released in March by the Government Accountability Office warned that “trust in real media could be undermined by false claims that real media are deepfakes.” In other words, AI makes it easier for fake news to trump real news.
The problem is growing rapidly. “Between 2019 and 2020, the number of deepfake online content increased by 900%,” according to the World Economic Forum.
But there are cures for the toxins that AI can generate.
At the hearing, Zohaib Ahmed, CEO and co-founder of Resemble AI in Santa Clara, California, advocated for “creating a public database in which all generated election content is recorded, allowing voters to “easily access information on the nature of the origin of the vote”. content they encounter. He and others have also suggested using digital watermarking technology to verify the authenticity of content.
Whatever remedies are used, the time has come. Some actions are already underway. In February, following a request from Klobuchar and Collins, the bipartisan U.S. Election Assistance Commission voted unanimously to authorize federal funds to combat misinformation “amplified by AI technologies.”
But “by the time deepfake becomes widespread, any reports calling it fake are also too late,” said Ben Colman, CEO and co-founder of Reality Defender, a tool that can detect deepfakes. “This is not an alarmist campaign, nor alarmism, catastrophism or conspiratorial hyperbole. This is simply the logical progression of the weaponization of deepfakes.
He applauded the legislation, the Protecting Elections from Deceptive AI Act, but called for more action “imposing real sanctions on bad actors” who “transform reality and on platforms that fail to stop their spread.
It’s a personal topic for Klobuchar, who has spoken out about a dishonest Russian photo showing her funding Nazis in Ukraine. “That photo had a red circle around me in the background,” she said, “and then they put the police signs back into the hands of the people at the rally who were never there.”
Klobuchar called for quick action on her bill and a strong approval vote in committee so that “we can get this thing heard right away…
“We really can’t wait.”
washingtonpost
