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.
Peloton operates AI-powered personalization features, including a product called Peloton IQ, that process fitness and health goal data to generate recommendations; the policy states that data shared with third-party AI service providers is deidentified before sharing and is contractually restricted from being used for general AI model training or purposes unrelated to Peloton.
This analysis describes what Peloton'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 establishes that health-adjacent and fitness performance data, including inferences derived from workout activity, feeds internal machine learning models and direct AI interactions under features such as Peloton IQ. The provision includes a stated opt-out mechanism but does not specify the scope of data categories that feed AI training after opt-out.
Interpretive note: The scope of data categories that continue to feed AI model training after a user opts out of 'Personalized Experiences' is not explicitly specified in the document, creating ambiguity about the practical effect of the opt-out on training data flows.
Under this clause, the agreement authorizes Peloton to use fitness performance data, health goals, and behavioral inferences to train internal AI models and to power AI features that interact directly with members. The policy states that data shared with third-party AI providers is deidentified and contractually restricted from external AI training, and that members can opt out of personalized experiences via the Preferences toggle.
Cross-platform context
See how other platforms handle AI Personalization and Health Goal Data Processing (Peloton IQ) and similar clauses.
Compare across platforms →Monitoring
Peloton has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 20 platforms.
"In some instances, we may develop AI powered features within our Products where you interact directly with AI which leverages your fitness and health goals in order to recommend plans and provide tips for improving performance (e.g. Peloton IQ). You always have control and can opt out by toggling off "Personalized Experiences" in the Privacy section of Preferences in your Profile. [...] Some of these Inferences based on activity and recommendations utilize internal machine learning models and may be supported by AI Service Providers. We use this data to build, train, analyze and improve the accuracy of our models. We manually review the recommendations and Inferences based on activity produced by the models against the underlying data from which the recommendations and/or personalizations were made.Excerpt from Peloton's Privacy Policy
REGULATORY LANDSCAPE: This provision implicates GDPR provisions on automated decision-making and profiling, particularly where AI-generated recommendations are based on health-adjacent fitness data that may constitute special category data under GDPR Article 9. The EU AI Act's risk classification framework may apply to AI systems that leverage health-related data for recommendations. CCPA and state health data statutes, including Washington's My Health MY Data Act, may classify fitness and health goal data as regulated health information depending on context and jurisdiction. The FTC has enforcement interest in AI systems that process sensitive consumer data. GOVERNANCE EXPOSURE: High. The use of fitness and health goal data to train internal machine learning models and to power direct AI interactions creates compliance exposure across multiple regulatory frameworks, particularly where health-adjacent data categories (pregnancy workouts, mood workouts, heart rate) are included in training data. The manual review process described does not constitute a formal human oversight mechanism as defined under the EU AI Act. JURISDICTION FLAGS: EU/EEA exposure is heightened where AI recommendations based on health-adjacent data may trigger GDPR Article 22 rights regarding automated decision-making. Washington State My Health MY Data Act exposure applies where fitness data related to health conditions is processed. Illinois BIPA considerations arise where physical movement and form analysis data feeds AI models. California CCPA sensitive personal information provisions may apply to health-related inferences. CONTRACT AND VENDOR IMPLICATIONS: The policy asserts that third-party AI service providers are contractually restricted from using shared data to train general models or for independent purposes. Procurement and vendor management teams should assess whether Peloton's vendor contracts with AI providers include enforceable data use restrictions, audit rights, and breach notification obligations consistent with GDPR Article 28 processor requirements and CCPA service provider requirements. COMPLIANCE CONSIDERATIONS: Legal teams should evaluate whether the opt-out mechanism for personalized experiences also stops data from being used in AI model training, or only stops personalized outputs from being surfaced to the user. Data mapping should distinguish between training data flows and inference output flows. The deidentification standard applied to data shared with third-party AI providers should be assessed against GDPR pseudonymization requirements and CCPA's definition of deidentified data, including whether technical and organizational safeguards prevent re-identification.
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 establishes that health-adjacent and fitness performance data, including inferences derived from workout activity, feeds internal machine learning models and direct AI interactions under features such as Peloton IQ. The provision includes a stated opt-out mechanism but does not specify the scope of data categories that feed AI training after opt-out.
Under this clause, the agreement authorizes Peloton to use fitness performance data, health goals, and behavioral inferences to train internal AI models and to power AI features that interact directly with members. The policy states that data shared with third-party AI providers is deidentified and contractually restricted from external AI training, and that members can opt out of personalized experiences …
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Peloton.