The San Francisco-based 9th U.S. Circuit Court of Appeals dealt Amazon a significant setback on Tuesday by reversing a lower court's temporary ban on Perplexity's artificial intelligence shopping tools. The decision represents a watershed moment in how American courts will approach the rapidly evolving landscape of agentic AI systems—autonomous software that can perform complex tasks including browsing the internet, making purchases and conducting transactions with minimal human input. For the technology industry and particularly for companies developing next-generation AI applications, the ruling signals that courts are inclined to permit user-directed AI agents to access commercial platforms even without explicit permission from the platform operator.
Amazon had initiated legal action against Perplexity in November, arguing that the AI startup was illicitly accessing customer accounts through its Comet browser and associated AI agent functionality. According to Amazon's complaint, Perplexity's system posed substantial security vulnerabilities and had continued unauthorised access despite repeated demands to cease operations. The technology allowed users to log into their Amazon accounts through Perplexity's AI interface, enabling the agent to browse products, review user preferences and place orders autonomously on behalf of customers. Amazon contended these activities violated the Computer Fraud and Abuse Act, a cornerstone US legislation regulating unauthorised computer access.
Perplexity disputed Amazon's characterisation entirely, arguing that the lawsuit lacked legal foundation and represented a deliberate attempt to prevent Amazon customers from utilising alternative AI tools. The company's defence highlighted a provocative perspective: Perplexity's AI agents, unlike human browsers, cannot perceive the advertising that Amazon saturates its platform with, thereby reducing Amazon's advertising exposure and commercial value. This counterargument framed the dispute not as a security issue but as a competitive challenge to Amazon's business model, suggesting the retail giant was using legal mechanisms to protect its advertising revenues rather than genuine security interests.
In March, a California federal judge initially sided with Amazon, issuing a temporary injunction that prohibited Perplexity from deploying its agentic AI on Amazon's platform. The judge determined that Amazon had presented substantial evidence demonstrating that Perplexity's AI agent constituted a violation of federal computer fraud statutes. This preliminary decision appeared to establish a boundary around how aggressively AI systems could interact with commercial platforms without explicit authorisation, potentially constraining the development of agentic AI tools across the technology sector.
The appeals court's Tuesday reversal fundamentally reframed the legal question. Rather than focusing on Perplexity's actions as a company, the 9th Circuit emphasised that it was the individual users employing Perplexity's AI agents who technically "accessed" Amazon's platform. This distinction matters profoundly because the Computer Fraud and Abuse Act targets unauthorised access by individuals, not software companies facilitating user access. By placing responsibility on the end user rather than the technology provider, the court created a conceptual framework that could permit sophisticated AI agents to operate across commercial websites provided they are acting on explicit user instruction and behalf.
This interpretation carries substantial implications for the emerging agentic AI sector. Companies developing AI tools designed to automate shopping, travel booking, financial management and similar transactional activities have operated in significant legal uncertainty. The 9th Circuit's ruling provides reassurance that user-directed AI agents may operate across commercial platforms without violating existing computer fraud legislation, provided the underlying user has authorised the agent's actions. This opens pathways for the proliferation of autonomous shopping assistants, travel planning agents and financial management tools that integrate directly with existing e-commerce and service platforms.
For Southeast Asian technology companies and investors monitoring American legal developments, the decision signals permissive regulatory boundaries around agentic AI. Jurisdictions across Asia are actively developing their own regulatory frameworks for artificial intelligence, and American court interpretations often influence subsequent policymaking in the region. If agentic AI agents receive legal validation in the world's largest technology market, regulatory authorities in Malaysia, Singapore, Thailand and other Southeast Asian nations may similarly embrace permissive approaches rather than restrictive ones.
Amazon responded with evident frustration, stating it remained confident in the underlying merits of its case and was evaluating subsequent legal options. The company's statement reflected a company confronting a novel landscape where traditional platform control mechanisms may prove insufficient against sophisticated AI agents. Amazon could pursue further appeals or pursue legislative remedies, attempting to persuade Congress to strengthen computer fraud statutes specifically addressing AI agent access. Alternatively, Amazon might implement technical barriers designed to detect and block AI agents accessing its platform, shifting the competition to the technical rather than legal domain.
Perplexity's spokesperson emphasised user choice and autonomy, framing the dispute as a broader principle about users' freedom to employ whatever artificial intelligence tools they prefer. This rhetorical position resonates with longstanding internet principles celebrating open access and user agency, positioning Perplexity's AI shopping agent as a consumer-empowerment tool rather than a security threat. The company's triumph in the appeals court validates this framing, at least in the eyes of federal judges.
The decision arrives at a moment when agentic AI development is accelerating substantially. Companies are investing billions in autonomous systems capable of planning, reasoning and executing complex tasks across digital ecosystems. Without clear legal boundaries, many technology companies have proceeded cautiously, uncertain whether their AI agents might trigger costly litigation or regulatory enforcement. The 9th Circuit's ruling removes substantial uncertainty, establishing that user-authorised AI agent access to commercial platforms does not inherently constitute unauthorised computer access under existing federal law.
However, the ruling does not entirely resolve all tensions between agentic AI development and platform operators' interests. Amazon and other major platforms retain substantial capacity to implement technical access controls, modify terms of service or pursue alternative legal theories to restrict unwanted AI agent activity. The decision establishes that simple unauthorised access claims may fail, but it does not preclude platforms from suing on other grounds including breach of contract, tortious interference or intellectual property violations. The litigation landscape surrounding agentic AI remains contested and evolving.
For platform operators globally, including in Southeast Asia, the American appeals court's reasoning suggests that broad bans on AI agent access may face legal challenges in jurisdictions with comparable computer fraud statutes. Technology companies planning agentic AI tools targeted at Asian markets can cite the 9th Circuit's reasoning in conversations with platform operators and regulators. The decision fundamentally shifts negotiating power toward AI agent developers and away from traditional platform gatekeepers, potentially accelerating the integration of autonomous AI systems into existing digital infrastructure throughout the Asia-Pacific region.
