ARTICLE
16 September 2026

Ninth Circuit Vacates CFAA Injunction Against Perplexity's Comet AI Agent

JD
Jones Day

Contributor

Jones Day is a global law firm with more than 2,500 lawyers across five continents. The Firm is distinguished by a singular tradition of client service; the mutual commitment to, and the seamless collaboration of, a true partnership; formidable legal talent across multiple disciplines and jurisdictions; and shared professional values that focus on client needs.
The Ninth Circuit Court of Appeals vacated a preliminary injunction against Perplexity AI, determining that when an AI agent acts at a user's direction, it is the user—not the AI company—who "accesses" computer systems under federal and California computer fraud statutes. This landmark appellate decision addresses how existing computer-access laws apply to agentic artificial intelligence and has significant implications...
United States Media, Telecoms, IT, Entertainment
Mary Alexander Myers’s articles from Jones Day are most popular:
  • within Media, Telecoms, IT and Entertainment topic(s)
  • in United Kingdom
  • with readers working within the Technology, Oil & Gas and Telecomms industries
Jones Day are most popular:
  • within Media, Telecoms, IT, Entertainment and Tax topic(s)

The Ninth Circuit Court of Appeals ("the Ninth Circuit") vacated a preliminary injunction against Perplexity AI, holding that Perplexity's artificial intelligence ("AI") agent acting at a user's direction does not independently "access" a computer system under the Computer Fraud and Abuse Act ("CFAA") or California's analogous statute.

Background

In a prior Alert, we described how a federal court entered a preliminary injunction after finding that Perplexity AI's agentic browser "Comet" likely violated the CFAA by accessing Amazon's systems without authorization after receiving a cease-and-desist letter, extending the Power Ventures framework to AI agents for the first time.

Computer System Access

On August 4, 2026, the Ninth Circuit vacated the preliminary injunction. Writing for the panel, Judge Milan D. Smith Jr., concluded that, based on the record before the court, it is the user—not Perplexity—who "accesses" Amazon's computers for CFAA purposes. The court's reasoning turned on how Perplexity's Comet agent actually functions: a user directs the agent through their own browser; the browser communicates with Amazon's servers while the agent takes screenshots of the browser view, sends them to Perplexity's servers for processing, and receives navigation instructions in return. At no point do Perplexity's servers directly communicate with Amazon's servers. Because the CFAA applies to "[w]hoever … intentionally accesses" a protected computer—language contemplating access by a person, not an AI tool—the court held that the user, not Perplexity, is the party who "accesses" the system. The panel also invoked the rule of lenity, reasoning that imposing liability on Perplexity would require a novel interpretation far afield from the statute's anti-hacking purpose and could expose individual users to criminal liability. The panel reached the same conclusion under the California Comprehensive Computer Data Access and Fraud Act ("CDAFA").

Significance

The opinion is the first appellate decision addressing how computer-access statutes apply to agentic AI. The court expressly limited its holding to the current record and acknowledged that the legal understanding of agentic AI will change as the technology evolves.

Companies deploying AI agents should evaluate this decision alongside their compliance strategies, while platform operators may need to rely on alternative enforcement mechanisms, including terms of service restrictions and technological measures, to regulate third-party AI access to their systems.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

[View Source]
See More Popular Content From

Mondaq uses cookies on this website. By using our website you agree to our use of cookies as set out in our Privacy Policy.

Learn More