OpenAI API customer content is not used to train models unless the customer opts in. Abuse-monitoring logs may retain prompts and responses for up to 30 days by default. Certain stateful features retain application data for 30 days or until deleted. Eligible approved customers may configure Modified Abuse Monitoring or Zero Data Retention, subject to endpoint limitations.
OpenAI data controlsResponsible AILegal
AI Policy
Future Mortgage uses artificial intelligence and automation to make the mortgage experience faster, clearer, and more consistent. We pair that technology with strong data safeguards, approved business systems, and qualified people who remain accountable for your loan.
01How we use AI
AI and automated tools may support document classification and extraction, workflow routing, quality checks, fraud and risk indicators, pricing comparisons, customer-service support, and drafting or summarizing information for review.
02Information AI systems may process
Depending on the workflow, an approved AI service may process application data, documents, communications, property information, financial information, identifiers, and technical metadata. We seek to limit each transfer to information reasonably necessary for the approved task and restrict access through company-controlled systems and credentials.
03Protecting sensitive document data
Before supported mortgage documents enter an approved AI workflow, our security controls identify sensitive values such as Social Security numbers and bank account numbers and mask all but limited trailing digits. For scanned or image-based documents, optical character recognition helps locate and protect the same sensitive information. Original documents remain in protected mortgage systems, while the AI workflow receives only the information reasonably needed for its approved task.
04Human oversight
AI supports our team; it does not replace the licensed professionals and authorized decision-makers responsible for reviewing applications, communicating with borrowers, and applying lending requirements. Material decisions remain subject to applicable law, program rules, data verification, and human oversight.
05Fair lending and responsible use
We design and monitor AI-supported processes with fair-lending, privacy, security, explainability, accuracy, and appropriate-use principles in mind. We do not permit AI to be used to unlawfully discriminate or bypass consumer-protection obligations.
06Future Mortgage retention
Future Mortgage retains mortgage application, transaction, communication, compliance, and related business records only for as long as reasonably needed to provide services and meet legal, regulatory, audit, fraud-prevention, security, dispute-resolution, and recordkeeping obligations. Deleting information from an AI provider does not necessarily delete the underlying record from Future Mortgage systems when retention is required.
07Vendor retention and exceptions
The periods below describe vendor-published standard practices for commercial or paid API services as of the last-updated date. Actual retention depends on the product, endpoint, feature, contract, account configuration, legal obligations, and misuse or security investigations. Stored files, conversations, caches, fine-tuning data, feedback, and stateful features may remain until deleted or for a separately stated period.
08Security and approved use
AI access is limited to company-approved systems, authorized people, and approved business purposes. Sensitive borrower information may not be entered into public or personal AI accounts. We combine role-based access, company-controlled credentials, sensitive-data masking, vendor security review, contractual protections, monitoring, and data minimization to help protect borrower information throughout the workflow.
09Questions or concerns
If you believe an AI-supported interaction was inaccurate or unfair, or you want a person to review an issue, contact care@futuremortgage.com or call 800.419.3630. Privacy questions may be sent to privacy@futuremortgage.com.
10AI provider data practices
ProviderStandard retention and training terms
Anthropic states that commercial API inputs and outputs are not used for model training by default and are automatically deleted from its backend within 30 days. Exceptions may apply for separately agreed retention, approved zero-data-retention arrangements, Usage Policy enforcement, legal requirements, saved product conversations, or content intentionally submitted as feedback.
Anthropic retention policyFor paid Gemini API services, Google states prompts and responses are not used to improve its products. Retention varies by feature: Google Search or Maps grounding retains prompts, context, and output for 30 days; Live API session-resumption state may remain for up to 24 hours; files remain until deleted or expired; explicit caches use their configured expiration; and implicit in-memory caches have a 24-hour TTL. Approved projects may use zero-data-retention controls, with feature limitations.
Gemini retention controlsVendor terms can change. Future Mortgage reviews its AI providers and configurations periodically and updates this notice when material practices change.