Welcome to this week’s issue of AI: The Washington Report, a joint undertaking of Mintz and its government affairs affiliate, ML Strategies.
The accelerating advances in artificial intelligence (“AI”) and the practical, legal, and policy issues AI creates have exponentially increased the federal government’s interest in AI and its implications. In these weekly reports, we hope to keep our clients and friends abreast of that Washington-focused set of potential legislative, executive, and regulatory activities.
This issue is the first of a three-part series of business guidance newsletters on regulatory trends in the AI space. Our key takeaways are:
- A lack of concrete regulation from Congress is not stopping consumer protection agencies from exercising their existing authority to regulate AI, as the Federal Trade Commission (“FTC” or “Commission”) has done in the Automators AI complaint.
- Businesses looking to offer AI-related products and services should be aware of the three primary concerns of consumer protection agencies with regard to AI: unfair or deceptive acts or practices, bias and discrimination, and anticompetitive conduct.
- To guard against allegations of unfair or deceptive conduct with regard to an AI product or service, businesses should conduct risk assessments of such products or services prior to launch, and strive for transparency, honesty, and fairness with their AI product or service.
Automators Case Signals Agencies’ Willingness and Ability to Regulate AI
As discussed in our previous newsletter, in early August the Federal Trade Commission (“FTC” or “Commission”) filed a first-of-its-kind individual case on AI-related misrepresentations against Automators AI (“Automators”). Consumer protection agencies have for months signaled their intent to utilize their existing statutory authority to regulate AI. The Automators complaint signals that these agencies were serious in their pronouncements.
AI presents businesses with significant opportunities to increase efficiency, cut costs, and offer innovative products and services. It is this transformative potential that makes AI a large and looming target for regulation.
In the wake of the Automators complaint, businesses looking to offer AI-related products and services should be aware of the three primary concerns of consumer protection agencies with regard to AI: unfair or deceptive acts or practices, bias and discrimination, and anticompetitive conduct.
Over the course of three separate newsletters, we will cover each of these concerns in turn, offering concrete and actionable guidance to businesses providing AI products and services. In this first of three newsletters, we discuss what such businesses need to know about the FTC’s intent to utilize its authority over “unfair or deceptive acts or practices” to regulate AI.
“Unfair or Deceptive Acts or Practices” in the AI Age
A bedrock component of the FTC’s enforcement authority derives from Section 5 of the Federal Trade Commission Act (“Section 5”). Section 5 grants the Commission the authority to prohibit “unfair or deceptive acts or practices in or affecting commerce…” In the past, the Commission has utilized its Section 5 authority to regulate fields for which it does not have explicit regulatory authority, such as online privacy for adults.
With the advent and popular adoption of generative AI tools, the FTC has released business guidance providing its interpretation of the regulatory authority over AI granted by Section 5. The FTC has divided business guidance on this topic between those AI-related business practices it would consider deceptive, and those it would consider unfair.
Deceptive Business Practices Related to AI
FTC guidance on AI-related deception discusses two distinct but interrelated business practices: false or misleading claims regarding AI products or services, and the leveraging of generative AI tools to mislead consumers.
Deception About AI
The alleged pronouncement of false or misleading claims about AI products or services lie at the heart of the FTC’s recent complaint against Automators. As covered in depth in our previous newsletter, the complaint alleges that Automators along with affiliated entities and individuals have caused individuals “over $22 million in harm” through false and misleading claims, some of which relate to the purported efficacy of the company’s AI tools.
While Automators allegedly claimed that it could help clients leverage generative AI tools like ChatGPT, “to scale an Amazon store to [$10,000] a month and beyond,” the FTC found that the majority of clients “do not recoup their investment, let alone make the advertised amounts…” For the Commission, these “false, misleading or unsubstantiated” earnings claims “constitute a deceptive act or practice in violation of Section 5(a)” of the FTC Act.
So as to not fall into the position of Automators and be the subject of an FTC complaint, it is important that businesses offering AI-related products and services take note of FTC business guidance on this topic. In February 2023, Michael Atleson of the FTC Division of Advertising Practices released a blog post entitled “Keep your AI claims in check.” In the article, Atleson warns those marketing their AI products or services “not to overpromise what your algorithm or AI-based tool can deliver.” Specifically, Atleson provides businesses with four recommendations that can help them avoid behaviors that would attract FTC scrutiny.
- Do not exaggerate the capabilities of your AI product.
- Do not baselessly assert that your AI product performs better than a given non-AI product.
- Duly consider and address “the reasonably foreseeable risks and impact of your AI product before putting it on the market.”
- Do not claim that a product uses AI when it in fact does not.
Deception Through AI
A distinct but related deceptive practice on the FTC’s radar is the use of generative AI tools to deceive, defraud, or mislead consumers. A March 2023 business guidance article entitled “Chatbots, deepfakes, and voice clones: AI deception for sale,” also authored by Atleson, provides context on how the FTC construes this type of business practice.
With the proliferation of AI powered chatbot, deepfake, and voice clone services, Atleson warns that the deployment of any AI tool that is “effectively designed to deceive – even if that’s not its intended or sole purpose” could be found to constitute a deceptive act or practice in violation in Section 5. Crucially, this means that the current FTC may pursue cases in which AI tools have the function of deceiving consumers, regardless of whether the tool was intended to deceive.
Atleson provides businesses offering generative AI products and services with four recommendations to avoid deceiving consumers in a manner that may violate Section 5.
- Consider whether the risks posed by your AI product or service are high enough to justify not bringing the product or service to market.
- Commit to effectively mitigating the reasonably foreseeable risks of your product or service prior to going to market.
- Do not overly-rely on post-release detection.
- Ensure that your AI product or service does not mislead people.
Unfair Business Practices Related to AI
Along with deceptive business practices, Section 5 of the FTC Act also gives the Commission regulatory authority over unfair business practices. A May 2023 article by Atleson entitled, “The Luring Test: AI and the engineering of consumer trust,” specifies that a practice is unfair “if it causes or is likely to cause substantial injury to consumers that is not reasonably avoidable by consumers and not outweighed by countervailing benefits to consumers or to competition.”
With regard to generative AI, Atleson asserts that the FTC is closely monitoring “design elements that trick people into making harmful choices,” causing people to “take actions contrary to their intended goals.” As a model for possible future actions by the Commission on unfair AI practices, Atleson points to previous FTC actions regarding allegedly unfair design choices for in-game purchases and mechanisms to cancel service.
To avoid falling under FTC scrutiny for unfair AI practices, Atleson recommends that firms provide clarity to users regarding whether generative AI content is organic or paid. Atleson also reiterates the importance of firms anticipating and responding to reasonably foreseeable risks posed by their AI products and services.
Conclusion: Steering Clear of Section 5 Violations
As the Automators case demonstrates, the FTC is willing and able to leverage its Section 5 authority to regulate providers of AI-related products and services. Enthusiasm surrounding the limitless business potential of AI tools should be tempered by a recognition of the resolve of agencies like the FTC to regulate AI, even in the absence of explicit instruction from Congress. Firms operating in this regulatory environment would do well to take seriously FTC pronouncements and business guidance surrounding AI.
While each business should consider its own circumstances and adjust accordingly, FTC guidance presents firms with two primary means by which the risk of costly regulatory scrutiny may be lessened.
- Consider the reasonably foreseeable risks of your AI product or service prior to releasing it to market. Address the risks accordingly. If such risks are too numerous or potentially harmful, consider not releasing the product or service in question.
- Strive for transparency, honesty, and fairness with your AI product or service. Do not leverage AI to deceive or manipulate consumers, or exaggerate the efficacy of your AI tool.
In subsequent editions of this series, we will consider how businesses can minimize the risk of regulatory scrutiny regarding bias or discrimination and anticompetitive conduct for their AI product or service.