terms of service

Subscribe to terms of service via RSS

In May 2024, we released Part I of this series, in which we discussed agentic AI as an emerging technology enabling a new generation of AI-based hardware devices and software tools that can take actions on behalf of users. It turned out we were early – very early – to the discussion, with several months elapsing before agentic AI became as widely known and discussed as it is today. In this Part II, we return to the topic to explore legal issues concerning user liability for agentic AI-assisted transactions and open questions about existing legal frameworks’ applicability to the new generation of AI-assisted transactions.

Background: Snapshot of the Current State of “Agents”[1]

“Intelligent” electronic assistants are not new—the original generation, such as Amazon’s Alexa, have been offering narrow capabilities for specific tasks for more than a decade. However, as OpenAI’s CEO Sam Altman commented in May 2024, an advanced AI assistant or “super-competent colleague” could be the killer app of the future. Later, Altman noted during a Reddit AMA session: “We will have better and better models. But I think the thing that will feel like the next giant breakthrough will be agents.” A McKinsey report on AI agents echoes this sentiment: “The technology is moving from thought to action.” Agentic AI represents not only a technological evolution, but also a potential means to further spread (and monetize) AI technology beyond its current uses by consumers and businesses. Major AI developers and others have already embraced this shift, announcing initiatives in the agentic AI space. For example:  

  • Anthropic announced an updated frontier AI model in public beta capable of interacting with and using computers like human users;
  • Google unveiled Gemini 2.0, its new AI model for the agentic era, alongside Project Mariner, a prototype leveraging Gemini 2.0 to perform tasks via an experimental Chrome browser extension (while keeping a “human in the loop”);
  • OpenAI launched a “research preview” of Operator, an AI tool that can interface with computers on users’ behalf, and launched beta feature “Tasks” in ChatGPT to facilitate ongoing or future task management beyond merely responding to real time prompts;
  • LexisNexis announced the availability of “Protégé,” a personalized AI assistant with agentic AI capabilities;
  • Perplexity recently rolled out “Shop Like a Pro,” an AI-powered shopping recommendation and buying feature that allows Perplexity Pro users to research products and, for those merchants whose sites are integrated with the tool, purchase items directly on Perplexity; and
  • Amazon announced Alexa+, a new generation of Alexa that has agentic capabilities, including enabling Alexa to navigate the internet and execute tasks, as well as Amazon Nova Act, an AI model designed to perform actions within a web browser.

Beyond these examples, other startups and established tech companies are also developing AI “agents” in this country and overseas (including the invite-only release of Manus AI by Butterfly Effect, an AI developer in China). As a recent Microsoft piece speculates, the generative AI future may involve a “new ecosystem or marketplace of agents,” akin to the current smartphone app ecosystem.  Although early agentic AI device releases have received mixed reviews and seem to still have much unrealized potential, they demonstrate the capability of such devices to execute multistep actions in response to natural language instructions.

Like prior technological revolutions—personal computers in the 1980s, e-commerce in the 1990s and smartphones in the 2000s—the emergence of agentic AI technology challenges existing legal frameworks. Let’s take a look at some of those issues – starting with basic questions about contract law.

The currents around the Communications Decency Act just got a little more turbulent as the White House and executive branch try to reel in the big fish of CDA reform.

On July 27, 2020, the Commerce Department submitted a petition requesting the FCC initiate a rulemaking to clarify the provisions of Section 230 of the Communications Decency Act (CDA). Unrelated, but part of the same fervor in Washington to “rein in social media,” the leaders of the major technology companies appeared before the House Judiciary Antitrust Subcommittee at a hearing yesterday, July 29, 2020, to discuss the Committee’s ongoing investigation of competition in the digital marketplace, where some members inquired about content moderation practices. Moreover, last month, a pair of Senators introduced the PACT Act, a targeted (but still substantial) update to the CDA (and other CDA reform bills are also being debated, including a bill to carve out sexually exploitative material involving children from the CDA`s reach).

President Trump signed an Executive Order today attempting to curtail legal protections under Section 230 of the Communications Decency Act (“Section 230” or the “CDA”). The Executive Order strives to clarify that Section 230 immunity “should not extend beyond its text and purpose to provide protection for those who purport to provide users a forum for free and open speech, but in reality use their power over a vital means of communication to engage in deceptive or pretextual actions stifling free and open debate by censoring certain viewpoints.”

Section 230 protects online providers in many respects concerning the hosting of user-generated content and bars the imposition of distributor or publisher liability against a provider for the exercise of its editorial and self-regulatory functions with respect to such user content.  In response to certain moderation efforts toward the President’s own social media posts this week, the Executive Order seeks to remedy what the President claims is the social media platforms’ “selective censorship” of user content and the “flagging” of content that does not violate a provider’s terms of service.

The Executive Order does a number of things. It first directs:

  • The Commerce Department to file a petition for rulemaking with the FCC to clarify certain aspect of CDA immunity for online providers, namely Good Samaritan immunity under 47 U.S.C. §230(c)(2). Good Samaritan immunity provides  that an interactive computer service provider may not be made liable “on account of” its decision in “good faith” to restrict access to content that it considers to be “obscene, lewd, lascivious, filthy, excessively violent, harassing or otherwise objectionable.”
    • The provision essentially gives providers leeway to screen out or remove objectionable content in good faith from their platforms without fear of liability. Courts have generally held that the section does not require that the material actually be objectionable; rather, the CDA affords protection for blocking material that the provider or user considers to be “objectionable.” However, Section 230(c)(2)(A) requires that providers act in “good faith” in screening objectionable content.
    • The Executive Order states that this provision should not be “distorted” to protect providers that engage in “deceptive or pretextual actions (often contrary to their stated terms of service) to stifle viewpoints with which they disagree.” The order asks the FCC to clarify “the conditions under which an action restricting access to or availability of material is not ‘taken in good faith’” within the meaning of the CDA, specifically when such decisions are inconsistent with a provider’s terms of service or taken without adequate notice to the user.
    • Interestingly, the Order also directs the FCC to clarify if there are any circumstances where a provider that screens out content under the CDA’s Good Samaritan protection but fails to meet the statutory requirements should still be able to claim protection under CDA Section 230(c)(1) for “publisher” immunity (as, according to the Order, such decisions would be the provider’s own “editorial” decisions).
    • To be sure, courts in recent years have dismissed claims against services for terminating user accounts or screening out content that violates content policies using the more familiar Section 230(c)(1) publisher immunity, stating repeatedly that decisions to terminate an account (or not publish user content) are publisher decisions, and are protected under the CDA. It appears that the Executive Order is suggesting that years of federal court precedent – from the landmark 1997 Zeran case until today – that have espoused broad immunity under the CDA for providers’ “traditional editorial functions” regarding third party content (which include the decision whether to publish, withdraw, postpone or alter content provided by another) were perhaps decided in error.

Epic Games, Inc. (“Epic”) is the publisher of the popular online multiplayer videogame Fortnite, released in 2017. In recent years, Fortnight has gained worldwide popularity with gamers and esports followers (culminating in July 2019 when a sixteen-year-old player won the $3 million prize for winning the Fortnite World Cup).  Players, in one version of the game, are dropped onto a virtual landscape and compete in a battle royale to survive.  In the real world, Epic recently survived its own encounter – not with the help of scavenged weapons or shield potions – but through its well-drafted end user license agreement (“EULA” or “terms”).

Earlier this month, the District Court for the Eastern District of North Carolina granted Epic’s motion to compel individual arbitration of the claims of a putative class action.  The action arose in connection with a cyber vulnerability that allowed hackers to breach user accounts. The court concluded that the arbitration provision contained in the EULA was enforceable in this case, even where a minor was the person who ultimately assented to the terms. (Heidbreder v. Epic Games, Inc., No. 19-348 (E.D.N.C. Feb. 3, 2020)).