The case could set federal precedent for AI regulation and deepfake accountability.
"Is this you?" That was the subject line of an email sent to North Carolina Sen. DeAndrea Salvador this summer.
The email included a link to an ad for energy-efficient home appliances sold by a subsidiary of Whirlpool in Brazil.
"In low-income communities in São Paulo, the average electricity bill cost represents 30% of their monthly income," Salvador appears to say in the ad. "This is when energy becomes a burden."
The problem? She never actually said that.
The video, which had already won a Grand Prix award at the Cannes Lions International Festival of Creativity, used artificial intelligence to manipulate Salvador's words, drawing from an unrelated TED Talk she gave in 2018.
"It's just really unfortunate to have so directly experienced the misuse of technology, particularly in such a big scale," Salvador, a Democrat representing Mecklenburg County, told The News and Observer.
The ad's award has since been revoked and the agency that created it has issued a public apology, but Salvador isn't done.
She sued Whirlpool and Omnicom (the parent company of the ad agency that produced the video) in federal court, alleging that they engaged in unfair and deceptive trade practices by falsifying her endorsement of their products.
The lawsuit, which was first reported by Courthouse News, is one of the first of its kind and could set meaningful precedent for AI abuses in the future.
"I could not find an instance, particularly with an elected official involved, that utilized deepfakes and AI technology in this type of way, commercially," Salvador said. "That was another big reason for me to want to push forward on this case.
"... It can happen to anybody and everybody."
Representatives for Omnicom and Whirlpool did not respond to a request for comment.
Supporter of using AI 'responsibly'
Salvador, a three-term state senator and recent MBA graduate of the Massachusetts Institute of Technology, is no stranger to AI.
Her degree in business analytics focused on deep learning, and she has spent over a year contributing to a national bipartisan working group on AI.
In the state legislature, Salvador has sponsored a variety of bills to regulate AI, including a ban on deepfakes in political campaigns and an initiative to create standards for labeling AI content.
And just this week, Gov. Josh Stein appointed Salvador to a new bipartisan oversight group on AI in state government.
"It's not necessarily a zero-sum game," she said. "For me, I believe that we can innovate responsibly."
State government agencies are already adopting AI technology in their everyday workflow, with the state treasurer's office even undergoing an extensive pilot program with OpenAl earlier this year.
But lawmakers -- both at the state and federal level -- have yet to adopt any meaningful guardrails on the usage of artificial intelligence.
Salvador's AI bills never made it to a hearing and, at the federal level, Congress even considered setting a 10-year ban on AI regulation by states.
Salvador's lawsuit, however, adds to a growing list of legal challenges to AI use that could bring change from the courts.
"The harm that has occurred to me initially has happened," Salvador said. "But what I do hope is that this example -- both from shedding a light on it and also the legal case -- could hopefully protect others."