Another state has joined the list of those that require clear disclosure of the use of artificial intelligence (“AI”) in political ads, joining others that have addressed concerns about deep fakes corrupting the political process. Michigan’s Governor Whitmer just signed a bill that adds Michigan to 4 other states (Texas, California, Washington, and Minnesota) that have enacted laws requiring the clear identification of the use of AI in political ads.  As many media companies are struggling with their policies on AI, and as the federal government has not acted to impose limits on the use of AI in political ads (see our posts here and here), it has been up to states to adopt rules that limit these practices.

The Michigan bill, H.B. 5141, applies to “qualified political advertisements” which include any advertising “relating to a candidate for federal, state, or local office in this state, any election to federal, state, or local office in this state, or a ballot question that contains any image, audio, or video that is generated in whole or substantially with the use of artificial intelligence.”  A companion bill, H.B. 5143, defines “artificial intelligence” as “a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments, and that uses machine and human-based inputs to do all of the following: (a) Perceive real and virtual environments. (b) Abstract such perceptions into models through analysis in an automated manner. (c) Use model inference to formulate options for information or action.”

H.B. 5141 requires that political advertisers who use AI must “State that the qualified political advertisement was generated in whole or substantially by artificial intelligence.”  For audio, the statement must be clearly audible, included at the beginning or end of the ad, and last for at least 3 seconds.  For video, the disclosure must be spoken clearly for at least 3 seconds at the beginning or end of the ad and appear for at least 4 seconds in text that is readable by the average viewer.

The Michigan law imposes these obligations primarily on the political advertiser – not the media company transmitting the ad to the public.  It exempts from liability broadcasters and cable and satellite providers if they are paid to broadcast a political message.   For online providers, the law’s language is not entirely clear, but it appears that they are also exempt if they have a clearly written policy statement that is disclosed to advertisers noting that there is a prohibition on the use of AI in political ads without proper disclosure. 

In connection with their own programming, the penalties set out in the law do not apply to radio or television broadcasting stations, including cable or satellite television operators, programmers, or producers that broadcast a qualified political advertisement or a communication generated in whole or substantially by artificial intelligence as part of a bona fide newscast, news interview, news documentary, or on-the-spot coverage of bona fide news. However, even in those programs, the program must clearly acknowledge in some clear manner, easily heard or read by the average listener or viewer, that the message was generated in whole or substantially by artificial intelligence and does not accurately represent the speech or conduct of the depicted individual.

Notably, in what appears to be a bid to observe First Amendment protections, the law does not apply the disclosure obligations to qualified political advertisements that constitute “satire or parody.”  This broad exemption would appear to invite arguments that many political ads are satire or parody and thus do not require an AI disclosure.  This certainly would stretch what seems to be the intent of the law, but watch for such arguments.

As we are entering an election season sure to be contested with many aggressive political ads, broadcasters and other media companies need to be on the alert for new legal developments such as this Michigan law.  In addition, as we warned in previous articles (see, for instance, this one), even without a ban on running these ads without disclosure, the ads can still pose legal liability under traditional legal concepts such as defamation.  As we have advised before, have your lawyer on speed dial as we enter this treacherous political season.