WASHINGTON (Reuters) – U.S. Senator Michael Bennet, a Democrat who is active in artificial-intelligence issues, wrote to leading tech firms on Thursday to urge them to label AI-generated content and limit the spread of material aimed at misleading users.
In the letter to the chief executives of OpenAI, which released ChatGPT, and its partner Microsoft among others, Bennet said it was especially important for Americans to know when AI was used to make political content.
“Fabricated images can derail stock markets, suppress voter turnout, and shake Americans’ confidence in the authenticity of campaign material,” he wrote. “Continuing to produce and disseminate AI-generated content without clear, easily comprehensible identifiers poses an unacceptable risk to public discourse and electoral integrity.”
Bennet’s letter was also addressed to top executives of Meta Platforms, Twitter, TikTok, Alphabet and others.
The fear has been that AI could be used to make high-quality fakes that not only confuse voters but create better scams or other images with malicious intent. Lawmakers in Congress – including Senate Majority Leader Chuck Schumer – have expressed interest in fast action to rein in the worst elements of AI but so far there has been no imminent sign of significant legislation becoming law.
Bennet has introduced a bill that would require political ads to disclose whether AI was used to create imagery or other content.
Bennet noted that some companies – including OpenAI and Alphabet’s Google – have begun marking some content as AI- generated but added that those companies’ policies are “alarmingly reliant on voluntary compliance.”
He asked the executives to answer a series of questions by July 31, including what standards or requirements they employ to identify AI content and how those standards were developed and audited to establish effectiveness. He also asked what happens to users who violate the rules.
Twitter, which is owned by billionaire Elon Musk, responded to a request for comment with a poop emoji. Microsoft declined comment while TikTok, OpenAI, Meta and Alphabet did not immediately respond to a request for comment.
(Reporting by Diane Bartz in Washington; Editing by Matthew Lewis)
