AI Regulation, the Election Bring Challenges for Advertisers

As November 5th creeps closer, it seems American voters are being promised another nail-biter, with several swing states polling within the margin of error for the presidential race. 

While there are numerous important races at play at all levels of government come Election Day, the race for the White House, as well as for control of the House and Senate, are the ones likely to determine the next several years of national AI policy. Although AI hasn’t quite proven to be a culture war issue, that doesn’t mean the partisan cracks won’t start showing soon. 

It’s not as if there hasn’t been any legislation and regulation around AI; however, compared to the EU and some other countries around the world, it’s been a pretty light touch. Sixteen states have passed legislation covering the use of AI in political advertising; in 2023, the White House issued an executive order directing federal agencies to develop new standards for AI safety and security (among other things), and California recently passed an AI transparency bill—so there’s been movement. 

But compared to such comprehensive legislation like the EU’s AI Act, it would seem the American approach is shaping up to be much more iterative and likely much more sectoral, both at the federal and state levels. 

While the wait for the dust to settle might extend well beyond November 5th, it’s worth taking a look at some developing AI policy issues that will matter to advertising.

Copyright

The U.S. Copyright Office announced an initiative in 2023 to begin examining the issues AI presented to current copyright law and is rolling out findings in a series of reports, the first of which was issued in July on Digital Replicas. 

Why it matters: The use of AI presents many new challenges for advertisers and agencies when it comes to copyright, perhaps none so pressing as the issue that U.S. copyright law currently does not allow copyrights to be issued to nonhuman-generated works. These series of reports are expected to lay the groundwork for federal legislation on the intersection of AI and copyright.

Unlike the topics below, copyright is mostly a federal issue, as the Copyright Act of 1976 preempts most state laws on copyright. 

Advertising category specific

As of August 2024, 16 states have passed (and several others have proposed) laws governing the disclosures required when AI is used in political advertising. In July, the Federal Communications Commission (FCC) issued a Notice of Proposed Rulemaking requiring AI disclosures on all political advertisements in TV and radio broadcasting. 

Why it matters: While it’s easy to understand why this was an area of focus in a high-profile election year, it’s also clear that this type of legislation could be developing cafeteria style—i.e., legislators and regulators will focus on whatever category of advertising they currently consider “risky” and craft new legislation for that category, a process to be repeated as new concerns arise in different categories. This type of approach to AI regulation in advertising could create a complex and uncertain environment for advertisers, potentially limiting innovation and hindering efforts to protect consumers. 

Content provenance

The content provenance issue is an area that dovetails with issues around watermarking and disclosures. They’re definitionally different things, but increasingly they’re talked about in the same context. 

Why it matters: California’s new legislation, the California AI Transparency Act, requires providers of generative AI systems to make available AI detection tools, offers AI users options to include visible disclosures that content is AI-generated, and include latent, machine-readable manifest information that discloses the origins of the disclosures in AI-generated content. 

What’s obviously concerning about this kind of state-level approach is that it appears to be shaping up similarly to the privacy law landscape in the United States: a state-by-state patchwork of laws on a topic that is better suited to a federal preemptive standard. 

Much like California was the first state to introduce a comprehensive privacy law with the CCPA, California is now also the first state to introduce this type of broad AI transparency and watermarking law. Other states will follow. 

It matters who wins the White House, the Senate and the House—but the surprise might be that it’s not actually a Republican vs. Democrat divide. Instead, what’s needed is a willingness for any new White House leadership and members of Congress to engage in AI issues in a thoughtful, bipartisan manner, carefully consider feedback from experts and stakeholders, and ideally be willing to craft strong federal preemptive standards.