Twenty tech companies working on artificial intelligence announced Friday that they have signed a pledge to ensure their software does not interfere with elections, including in the United States.
Signatories range from tech giants like Microsoft and Google to small startups that allow people to create fake voices. Such generative AI products could be exploited to create convincing deepfakes of candidates in elections.
The agreement effectively recognizes that their products pose significant risks in a year when 4 billion people are expected to vote in elections around the world.
“Deceptive AI election content can mislead the public in ways that jeopardize the integrity of the electoral process,” the document says.
The agreement is also a recognition that lawmakers around the world are not responding quickly to rapid advances in generative AI, leaving the tech industry to seek self-regulation.
“As society reaps the benefits of AI, we have a responsibility to help ensure these tools are not weaponized in elections,” Microsoft Vice Chairman and President Brad Smith said in a statement.
The 20 companies signing the pledge include Adobe, Amazon, Anthropic, Arm, Eleven Labs, Google, IBM, Inflection AI, LinkedIn, McAfee, Meta, Microsoft, Nota, OpenAI, Snap, Stability AI, TikTok, TrendMicro, Truepic, It’s X. .
The agreement is voluntary and falls short of a complete ban on AI content in elections, as some have called for. The 1,500-word document lists eight steps companies are taking this year. This step includes developing new tools to distinguish between AI-generated images and authentic content, as well as being transparent with the public about notable developments.
Free Press, an advocacy group that supports an open internet, said the pledge amounts to an empty promise as tech companies backtrack on previous promises about election integrity after the 2020 election. The group called for more oversight by human judges.
“Voluntary commitments like those announced today are not enough to address the global challenges facing democracies,” Free Press senior adviser Nora Benavidez said in a statement.
“Each election cycle, tech companies promise to uphold vague democratic standards, only to fail to fully deliver on those promises,” she says.
Representative Yvette Clark (New York) welcomed the technology agreement and said she hopes Congress will build on it. Clark has sponsored legislation to regulate deepfakes and AI-generated content in political advertising.
“This could be a defining moment for this Congress, and the only unifying issue on which we can come together to protect this country and future Americans,” she said in a statement.
This year is said to be the biggest year in the history of democracy, with elections held in seven of the world’s 10 most populous countries. In addition to the US presidential election in November, nationwide voting is also scheduled in India, Russia and Mexico. Elections have already been held in Indonesia, Pakistan and Bangladesh this year.
Ahead of January’s New Hampshire primary, a fake robocall claiming to be from President Joe Biden sparked widespread alarm about the potential for fake audio, images and video in politics. This month, the Federal Communications Commission voted to outlaw robocalls that include AI-generated audio.
Individual technology companies are taking their own measures. Meta, which owns Facebook and Instagram, said it would try to label images created by AI, but warned last month that it did not have the technical capacity to do similar labels for AI-generated audio and video. did.
Nick Clegg, Meta’s global president, said the pledge was intended to be a “meaningful step from the industry” towards cracking down on deceptive content. But he said the effort also required efforts from governments and civil society.
“With so many big elections happening this year, it’s critical we do everything we can to prevent people from being fooled by AI-generated content,” he said in a statement.
The companies announced the agreement at the Munich Security Conference, an annual event where world leaders discuss a variety of issues. Vice President Kamala Harris and Israeli President Isaac Herzog will be attending this weekend’s meeting.
At the World Economic Forum held in Davos, Switzerland in January, the topic of generative AI dominated public and private discussions.
