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 discloses that Betterment shares customer personal information with generative AI service providers, including large language model vendors, and states that Betterment receives assurances from those providers that the data is not used to train their models.
This analysis describes what Betterment'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 discloses a category of third-party data processor not commonly detailed in financial services privacy policies and relies on vendor-provided assurances rather than specifying the contractual or auditable mechanisms that enforce the model training restriction.
Interpretive note: The enforceability and scope of the model training assurances are not described in the document, making it unclear whether they constitute binding contractual restrictions or informal representations.
The updated policy discloses a new Fully Paid Securities Lending program through Apex Clearing, under which Betterment will share customer personal information and account details with Apex if customers choose to participate. The revised terms also establish that generative AI service providers have committed that personal information will not be used for model training. For customers participating in promotional offers requiring offline fulfillment, the policy now explicitly states that personal information including mailing address may be shared with third-party partners. You can review the FPSL Program supplemental disclosures for details about the securities lending arrangement, or choose not to participate in the program.
View change record →The agreement authorizes sharing of personal information with generative AI vendors and large language model providers as part of Betterment's service delivery infrastructure. The protection against model training use rests on assurances received from those vendors, the contractual enforceability and audit rights of which are not described in the policy.
Cross-platform context
See how other platforms handle Generative AI Vendor Disclosure and Model Training Assurance and similar clauses.
Compare across platforms →Monitoring
Betterment has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 20 platforms.
"These service providers also include providers of generative artificial intelligence technologies (including large language models), from which we receive assurances that your personal information is not used for the purpose of model training.Excerpt from Betterment's Privacy Policy
1) REGULATORY LANDSCAPE: This provision engages FTC Act Section 5 authority over unfair or deceptive practices, as well as emerging regulatory guidance from the FTC on AI data practices. For California residents, CCPA/CPRA requirements regarding disclosure of service provider data processing limitations may apply. SEC and FINRA guidance on the use of AI in investment advisory contexts is also relevant given Betterment LLC's registered investment advisor status. The provision does not specify whether the AI vendors qualify as 'service providers' under CCPA, which would require a contract prohibiting the vendor from using data for its own purposes including model training. 2) GOVERNANCE EXPOSURE: High. The policy asserts a material protection (no model training) but grounds it solely in received assurances rather than disclosed contractual terms, audit rights, or certification mechanisms. If those assurances are not contractually binding or verifiable, the stated protection may not be operationally effective. 3) JURISDICTION FLAGS: California (CCPA/CPRA service provider requirements), Colorado (CPA), Virginia (VCDPA), and EU/EEA (GDPR Article 28 processor agreement requirements) create heightened exposure. GDPR would require a documented data processing agreement with AI vendors specifying permissible processing purposes, which this policy does not confirm exists. 4) CONTRACT AND VENDOR IMPLICATIONS: Procurement and legal teams should assess whether Betterment's agreements with AI vendors include explicit contractual prohibitions on model training using customer data, whether those agreements include audit or certification rights, and whether the vendors are appropriately classified as service providers or processors under applicable privacy law. The current disclosure does not confirm the existence or scope of such agreements. 5) COMPLIANCE CONSIDERATIONS: Compliance teams should evaluate whether the AI vendor disclosure satisfies applicable state privacy law service provider disclosure requirements, review data processing agreements with all named and unnamed generative AI vendors, assess whether data minimization practices are applied before sharing with AI providers, and determine whether customer consent or additional disclosure is required under applicable law for this category of data sharing.
This provision discloses a category of third-party data processor not commonly detailed in financial services privacy policies and relies on vendor-provided assurances rather than specifying the contractual or auditable mechanisms that enforce the model training restriction.
The agreement authorizes sharing of personal information with generative AI vendors and large language model providers as part of Betterment's service delivery infrastructure. The protection against model training use rests on assurances received from those vendors, the contractual enforceability and audit rights of which are not described in the policy.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Betterment.