The 9th U.S. Circuit Court of Appeals in San Francisco has delivered a landmark decision that fundamentally reshapes how courts view the legal boundaries of artificial intelligence agents operating on commercial platforms. In a ruling issued on Tuesday, the appeals panel overturned a previous injunction that had prohibited Perplexity, a San Francisco-based AI startup, from deploying its sophisticated shopping tools on Amazon's platform. The decision represents a watershed moment for the rapidly expanding field of agentic artificial intelligence, which encompasses systems capable of autonomous browsing, purchasing decisions, and transaction completion with minimal human intervention.
The core dispute centred on whether Perplexity's AI agents violated the Computer Fraud and Abuse Act, a foundational U.S. law governing unauthorized computer access. Amazon had initiated legal proceedings in November, claiming that Perplexity's Comet browser and associated AI agent were surreptitiously penetrating private customer accounts, executing purchases without explicit authorization, and creating security vulnerabilities. The e-commerce giant maintained that Perplexity had systematically ignored cease-and-desist demands to halt the practice. A federal judge in California initially sided with Amazon's position in March, issuing a temporary restraining order based on what the court termed "strong evidence" of legal violation.
The appellate court's reasoning hinged on a technical but consequential distinction regarding who constitutes the actual "accessor" under the hacking statute. Rather than viewing Perplexity as the entity gaining unauthorized entry, the 9th Circuit determined that the platform access occurred through Perplexity's users operating the AI agents on their own behalf. This interpretation effectively shields the company from direct liability under the Computer Fraud and Abuse Act, as the users themselves—not Perplexity as an intermediary—held legitimate authority to access their own accounts. The ruling's implications extend far beyond this single dispute, establishing precedent for how courts may evaluate automated agents acting under user delegation across the broader digital ecosystem.
Amazon responded to the appellate decision with measured resistance, issuing a statement indicating the company remains unconvinced by the court's analysis. A company spokesperson emphasized that Amazon "respectfully disagrees" with the preliminary injunction ruling and signalled continued confidence in the underlying legal theory, while announcing that management is deliberating over potential subsequent appeals or alternative litigation strategies. The statement suggests Amazon views this reversal as a temporary setback rather than a definitive conclusion to the broader dispute, maintaining faith that higher-level review or revised legal arguments might ultimately vindicate its position.
Perplexity responded with a public assertion of vindication, framing the decision as confirmation of consumer rights in selecting artificial intelligence tools. Company spokesperson Jesse Dwyer released a statement emphasizing that Perplexity views itself as defending users' prerogatives to choose preferred AI platforms, casting the case as part of a broader struggle against erosion of consumer choice. The characterization implicitly positions Perplexity as defending internet freedom against what it perceives as Amazon's attempt to monopolize access to its own ecosystem through aggressive legal action.
Amazon's original complaint articulated a provocative secondary argument regarding the economics of online retail. The company contended that Perplexity's AI agents, lacking visual perception of digital advertising, effectively enable users to circumvent the advertising infrastructure that underpins Amazon's business model. Perplexity dismissed this assertion, characterizing Amazon's lawsuit as a transparently strategic effort to eliminate competition from AI-mediated shopping experiences, specifically targeting the value proposition that AI agents provide by filtering out the "pervasive advertising" that Amazon employs to shape customer purchasing behaviour.
For Malaysian and Southeast Asian technology stakeholders, this decision carries substantial implications. As artificial intelligence adoption accelerates across the region's e-commerce and digital services sectors, questions about how national and international legal frameworks apply to autonomous agents become increasingly urgent. The 9th Circuit's precedent may influence how similar disputes are resolved in jurisdictions ranging from Singapore to Indonesia, where digital commerce plays an expanding economic role. Regional regulators and technology companies will likely monitor further developments in this case, as it establishes foundational principles about AI agent legitimacy that could shape competitive dynamics across multiple markets.
The broader significance of this ruling extends to the governance of agentic AI systems, which represent a technological paradigm distinct from conventional software. Unlike traditional applications that require explicit user interaction for each operation, agentic systems possess the capacity to plan, reason, and execute multi-step procedures with constrained human oversight. This autonomy creates novel legal and security questions that existing legislative frameworks were not designed to address. The court's decision implicitly acknowledges that assigning liability for agent actions presents theoretical challenges—if agents access systems on behalf of users, are users or developers responsible for ensuring proper authorization and security compliance?
The decision leaves unresolved numerous practical security concerns that motivated Amazon's litigation. Even if Perplexity's technical architecture does not violate the Computer Fraud and Abuse Act, questions persist about data protection, account security, and whether users fully understand what permissions they grant when deploying AI agents within their accounts. The appeals court focused narrowly on statutory interpretation rather than addressing these broader policy dimensions, potentially leaving regulatory gaps that alternative legal theories—including those premised on consumer protection statutes or contract law—might eventually address.
Looking forward, the ruling may catalyze legislative responses designed to create clarity around agentic AI operations. Congress, state legislatures, and international regulatory bodies may determine that existing computer fraud statutes inadequately address the operational realities of AI agents, prompting new legislation that directly addresses these systems. Such legislative development could, paradoxically, provide more certainty than the current legal grey zone, offering clear rules that technology companies and platforms can follow with confidence.
The dispute also illuminates competitive anxieties within the artificial intelligence industry. Amazon, increasingly active in developing its own AI capabilities, may view Perplexity not merely as a technical threat but as a competitive encroachment into the customer relationship that Amazon has cultivated through investments in platform infrastructure and user data accumulation. Perplexity's technological approach—enabling users to leverage AI for shopping decisions—represents a potential shift in how consumers interact with e-commerce, potentially reducing Amazon's ability to influence purchasing through algorithmic recommendations and advertising placement.
As agentic AI systems proliferate, similar disputes will inevitably emerge across multiple sectors and jurisdictions. Financial services, healthcare, travel, and other industries where automated agents could legitimately execute transactions face comparable questions about authorization, accountability, and platform access. The 9th Circuit's decision provides initial guidance, but the legal landscape remains unsettled, suggesting that technology companies, platforms, and regulators will continue wrestling with appropriate governance frameworks for years to come.
