Automation assists.
People remain responsible.

AI may assist people, but should not obscure responsibility. Use must be purposeful, proportionate, explainable to affected users, respectful of privacy and intellectual property, and subject to meaningful human control.

Approved tools may support research, classification, translation, document extraction, drafting, accessibility and workflow guidance where the user is told what the tool does and its output is checked before reliance or submission.

• no fabricated authority, evidence, credentials or professional conclusions;
• no undisclosed solely automated decisions with legal or similarly significant effects;
• no upload of confidential, special-category, passport or case data to an unapproved tool;
• no impersonation, manipulation, unlawful discrimination or circumvention of safeguards; and
• no use of output as a substitute for qualified advice where such advice is required.

The responsible department must appoint a person to check material facts, sources, context, bias, confidentiality, legal requirements and intended use before relying on AI-assisted work. Users must have an opportunity to confirm extracted or generated information before it is adopted into a submitted document or case record.

USG Growth remains responsible for its advisory outputs; DRP for its case-support and professional procedures; USG Academy for teaching, assessment and certification processes; USG Commons for its editorial and participation activities; and USG Partnerships for its appointment processes. AI does not replace the authority or responsibility of the person making a decision.

Before enabling an AI supplier for a particular use, USG must assess data roles, permitted inputs, training use, retention, security, sub-processors, transfer locations, deletion, intellectual-property terms and available controls. The approved use and any customer restrictions must be documented.

Confidential documents, identity material, special-category information and case data must not be entered into an unapproved tool. Customer or case content must not be used to train a general model without an appropriate legal basis, clear authority and transparent terms. A provider’s general marketing statement does not establish the settings or contractual protections of a particular USG account.

Material AI assistance must be explained at the point of use in a way relevant to the task, including its purpose and important limitations. Where AI-generated teaching or other media might reasonably be mistaken for a real person’s participation, its use should be made clear.

Affected users can ask the responsible department to clarify an output, correct source information and arrange human review of a relevant service decision. The reviewer must have the authority and information needed to reconsider the matter. The conflict and review safeguards in Conflicts of Interest and Complaints and Review apply.

The relevant service owner is accountable for approval, permissions, testing, monitoring, incidents and withdrawal of the AI use. Technical staff implement approved controls; the Privacy and Compliance functions review data-protection and governance implications. Permitted activities in this policy are not a statement that every listed feature is active in every service.

Governance and misuse concerns: compliance@universalsustainablegrowth.org. Data concerns: privacy@universalsustainablegrowth.org. Formal complaints: complaints@universalsustainablegrowth.org.