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 authorizes Verizon to use automated processing including artificial intelligence and machine learning to train algorithmic models using customer data for network management, marketing personalization, and service prediction. The policy includes a specific statement that customer personal information is not collected, used, or sold to train large language models.
This analysis describes what Verizon'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 customer data is used to train machine learning models for marketing and service prediction purposes, and includes an explicit carve-out stating that personal information is not used for large language model training, a disclosure that addresses a specific area of consumer and regulatory concern regarding generative AI.
Interpretive note: The operational scope of machine learning model training, including which customer data categories are used, which models are trained, and whether model outputs are used in automated decisions affecting individual customers, is not fully described in the policy text.
The agreement authorizes Verizon to train machine learning models using customer data for purposes including marketing personalization and service prediction. The policy explicitly states that personal information is not collected, used, or sold for large language model training.
Cross-platform context
See how other platforms handle AI and Predictive Analytics Use and similar clauses.
Compare across platforms →Monitoring
Verizon has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"We may use automated processing, including artificial intelligence, to assist us with these uses. For example, we use predictive analytics to monitor our network for faults and outages which helps us proactively and rapidly fix issues, and we may use machine learning techniques to train algorithmic models with information we have about you and other customers to provide more relevant marketing messages or predict what services might be of interest. We do not collect, use or sell personal information for the purpose of training large language models.Excerpt from Verizon's Privacy Policy
(1) REGULATORY LANDSCAPE: This provision engages the FTC Act's commercial surveillance and AI accountability frameworks, the EU AI Act (for any Verizon operations or customers within the EEA), and California's draft AI transparency and automated decision-making regulations under CPRA. The FTC has issued guidance on AI and automated decision-making practices, emphasizing that algorithmic model training using consumer data requires adequate disclosure and appropriate consent mechanisms. The specific LLM training exclusion is responsive to emerging regulatory attention on AI training data practices. (2) GOVERNANCE EXPOSURE: Medium. The authorization to train algorithmic models using individual customer data including browsing, location, and demographic data for marketing purposes creates compliance exposure under state automated decision-making frameworks. The policy does not describe the categories of models trained, retention periods for training data, or mechanisms for customers to contest decisions made by trained models, which may be required under emerging state AI accountability regulations. (3) JURISDICTION FLAGS: California's CPRA regulations include provisions addressing automated decision-making that may require opt-out rights for profiling used in marketing contexts. Colorado's privacy law includes automated decision-making protections. The EU AI Act categorizes certain automated profiling systems used in marketing as limited-risk or prohibited systems depending on their design, but this policy does not address EU operations. (4) CONTRACT AND VENDOR IMPLICATIONS: The use of customer data to train algorithmic models shared with or operated by service providers creates vendor assessment obligations. Legal teams should confirm whether model training involves disclosure of personal data to third-party AI vendors and whether applicable data processing agreements address model training as a permitted processing purpose. (5) COMPLIANCE CONSIDERATIONS: Legal teams should evaluate whether automated decision-making disclosures satisfy current and emerging state requirements, whether customers have accessible opt-out rights for profiling-based marketing that uses trained models, and whether the LLM training exclusion is operationally implemented and auditable. Data mapping should identify which customer data categories are used as training inputs and for which model types.
This provision establishes that customer data is used to train machine learning models for marketing and service prediction purposes, and includes an explicit carve-out stating that personal information is not used for large language model training, a disclosure that addresses a specific area of consumer and regulatory concern regarding generative AI.
The agreement authorizes Verizon to train machine learning models using customer data for purposes including marketing personalization and service prediction. The policy explicitly states that personal information is not collected, used, or sold for large language model training.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Verizon.