Get the weekly research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean. No account.
The policy prohibits using Perplexity's Services to make automated decisions that materially affect individuals in the domains of employment, healthcare, finances, housing, or insurance, or in contexts where errors could affect critical infrastructure or product safety.
This analysis describes what Perplexity AI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
This provision directly restricts high-value enterprise deployment scenarios in regulated industries and requires organizations to assess whether existing or planned Perplexity integrations fall within the prohibited categories before deployment.
Interpretive note: The provision does not define 'materially impact' or specify whether human-in-the-loop review removes a workflow from its scope, creating ambiguity for hybrid decision-making architectures.
Under this clause, users and organizations may not deploy Perplexity outputs to drive automated decisions affecting a person's employment status, healthcare treatment, financial standing, housing eligibility, or insurance coverage. The agreement establishes these restrictions across both consumer-facing and enterprise contexts.
Cross-platform context
See how other platforms handle Automated Decision-Making Prohibition and similar clauses.
Compare across platforms →Monitoring
Perplexity AI has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 20 platforms.
"Make automated decisions about individuals that materially impact them (e.g., decisions regarding their employment, healthcare, finances, housing or insurance) or where errors could be dangerous (e.g., decisions related to the management of critical infrastructure or safety components of products);Excerpt from Perplexity AI's Perplexity Acceptable Use Policy
(1) REGULATORY LANDSCAPE: This provision directly engages the EU AI Act's high-risk AI system classifications, which include employment, healthcare, credit, and essential services decision-making systems. It also engages the FTC's guidance on algorithmic decision-making and unfair or deceptive practices. GDPR Article 22, which governs automated decision-making with legal or similarly significant effects, is also implicated for EU/EEA-based deployments. Relevant enforcement authorities include the European AI Office, national data protection authorities, and the FTC. (2) GOVERNANCE EXPOSURE: High. Organizations currently using Perplexity in recruitment screening, insurance underwriting, credit scoring, clinical decision support, or housing eligibility workflows may be in breach of this provision as written. The prohibition is stated broadly without a de minimis threshold, and the enumerated categories align closely with regulated high-risk AI domains under the EU AI Act. (3) JURISDICTION FLAGS: Heightened exposure exists for EU/EEA organizations subject to the EU AI Act's high-risk system requirements, and for US organizations operating in financial services, healthcare, or employment contexts subject to FTC oversight and sector-specific regulation. Illinois, New York, and California have enacted or proposed legislation addressing automated employment decision tools, which may intersect with this provision. (4) CONTRACT AND VENDOR IMPLICATIONS: Enterprise customers and API integrators should audit their deployment architectures to determine whether Perplexity outputs feed into automated decision pipelines in the prohibited categories. B2B contracts that authorize such use cases may conflict with this AUP provision, and procurement teams should assess whether downstream liability flows to their organization. (5) COMPLIANCE CONSIDERATIONS: Compliance teams should conduct a use-case inventory of all Perplexity integrations to identify any deployment that generates outputs used in automated employment, healthcare, financial, housing, or insurance decisions. Policy updates and internal governance documentation should reflect this contractual restriction. Organizations should evaluate whether human-in-the-loop review mechanisms are sufficient to remove a workflow from the automated decision-making category as defined in this provision.
Regulatory citations, enforcement risk, and due diligence action items.
Provision-level monitoring, governance timelines, and regulatory mapping built from archived source documents and historical version tracking.
This provision directly restricts high-value enterprise deployment scenarios in regulated industries and requires organizations to assess whether existing or planned Perplexity integrations fall within the prohibited categories before deployment.
Under this clause, users and organizations may not deploy Perplexity outputs to drive automated decisions affecting a person's employment status, healthcare treatment, financial standing, housing eligibility, or insurance coverage. The agreement establishes these restrictions across both consumer-facing and enterprise contexts.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Perplexity AI.