AI Optimization?

September 5, 2024 

Privacy Plus+ 

Privacy, Technology and Perspective

This week, let’s focus on the challenges associated with artificial intelligence (AI) in advertising.

Background (From Guerilla Marketing to Content Creation):

Back in the early 2000s, guerilla marketing was all the rage. It thrived by using unconventional methods to promote products or services to target audiences. 

At the time, one of us was interning for a prominent advertising agency, working on a grapefruit account, among other things. The news had just broken:

“Grapefruit juice can kill old men with high blood pressure.”

So, it was time for a big rebranding of grapefruit. The account team decided that we “creatives” should target an audience of healthy and sassy women.

At this point, someone had a great guerilla marketing idea: “Why not place grapefruit front and center on a television show like ‘Sex and the City?’ Instead of cosmos, the ‘girls’ could drink grapefruit juice!”

Lo-and-behold the Ruby Martini—a mix of grapefruit juice and spirits—went mainstream in a certain demographic around 2002.

Years passed. The digital age arrived in full.

Eventually, guerilla marketing laid the groundwork for content creation, where advertisers pay “creators” to use their influence over a particular digital community to “create authentic brand interactions” to “drive engagement” – all jargon for “influencer” advertising, which essentially involves individuals marketing products or services to their “followers” online, often without disclosing their compensation.

The FTC’s Influencer Guidelines:

We've written before about the Federal Trade Commission’s (FTC) role in regulating the use of endorsements and testimonials in advertising, including its update to the Guides Concerning the Use of Endorsements and Testimonials in Advertising (16 C.F.R. Part 255) in 2022. In addition to containing a broader reflection around the historical use of endorsements, that post illuminates the Guides’ requirements. Generally, endorsements should be honest; reflect the actual experiences of the endorsers; experiences should be typical; and all endorsements should “clearly and conspicuously” disclose material connections.  Put another way, “#ad #sponsored [Post follows]. You can read that post by clicking on the following link:

https://www.hoschmorris.com/privacy-plus-news/endorsements-and-testimonials

The Guides have been brought up to date with a Final Rule, dated July 26, 2023. You can read the current Guides by clicking on the following link:

https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B/part-255

AI Optimization

Now comes the next evolution of advertising: The use of AI.

The use of AI in advertising takes several forms. Here, let’s focus on what’s being termed “AI optimization” (A.I.O.).

Currently, A.I.O. is being perceived as the successor to search engine optimization (S.E.O.), which developed as businesses sought to improve their rankings on Google and other search engines. Now, businesses are seeking to boost brands by influencing AI outputs.

Specifically, businesses are now paying to promote products and services by having those products and services placed in AI systems’ responses to consumers’ queries (especially through chatbots). 

So, as Kevin Roose recently explained in the NYT: “[C]ompanies are eager to insert themselves into chatbot responses, so that when a ChatGPT or Gemini user asks, ‘What’s the best restaurant in Dallas?’ or ‘Which S.U.V. should I buy?’ the chatbot recommends their products.” A link to his informative article follows:

https://www.nytimes.com/2024/08/30/technology/ai-chatbot-chatgpt-manipulation.html 

Our Thoughts

The core of false advertising law is consumer protection: consumers should not be misled on matters that would reasonably be expected to influence their purchasing decisions. 

To us, “influencing” generative AI systems (think: ChatGPT, Claude, Bard, Microsoft Copilot, Perplexity, etc.) so that they promote or guerilla market certain products or services to target audiences—those who query the AI system or chatbot with particular questions—amounts to bought-and-paid-for ad placement that skates along the knife-edge of fair and unfair advertising practices (or crosses directly over it). 

Today, AI systems and chatbots are embedded in modern technologies, often unseen and undisclosed. But even when you interact with one of these systems knowingly, the systems themselves seem like a helpful friend. Accordingly, suggestions by these systems—and their influence—can even seem very close to undisclosed paid endorsements, which would contravene the spirit of Section 5 of the FTC Act, and the letter of the Guides.

Indeed, if the bought-and-paid-for answers say or imply that they have assessed a variety of independent sources to find the answers to objectively provable questions, they would step on the line of outright fraud – even if they turn out to be correct (in which case the fraud would lie in having stated or implied that assessment had been made, when it had not).

We would suggest that the use of A.I.O. in these ways should be disclosed, and we believe that the FTC will one day state this expressly. Sponsored social media posts and S.E.O. results must tell consumers they are “sponsored.” In our view, sponsored A.I.O. outputs shouldn’t be any different.

Such disclosure will likely be required both by the companies utilizing A.I.O. and those owning the AI systems that have been influenced to influence their users. Without transparency around A.I.O., it’s hard to see how companies can avoid the consequences of undisclosed endorsements and testimonials, if not outright false advertising.

---

Hosch & Morris, PLLC is a boutique law firm dedicated to data privacy and protection, cybersecurity, the Internet and technology. Open the Future℠.