**Elevated Policy Note** **Subject:** Hierarchical Accountability Framework for Generative AI Access and Use **Core Premise** Access to, and responsible use of, generative AI should be governed by the natural “parent” authority closest to the individual. Accountability flows upward from the unit that has the clearest duty of care and the greatest practical ability to set norms, monitor use, and correct misuse. ### 1. Primary Accountability Layers | Layer | “Parent” Authority | Scope of Responsibility | Accountability Holder | |-------|--------------------|--------------------------|-----------------------| | Family / Household | Parents or legal guardians | Minors and dependent young adults living under parental care | Head(s) of household | | Educational Institution | School, college, university | All enrolled students and staff while acting in institutional capacity | Head of institution (principal, vice-chancellor, board) + academic governance bodies | | Organisation | Employer, professional body, membership association, NGO, religious body, etc. | Employees, members, or affiliates when using AI in organisational context or on organisational systems | Chief executive / governing board | | Local / Public Authority | Municipal or district administration | Individuals who fall outside the three layers above | Designated local AI coordination officer or equivalent public office | This creates a nested, not competing, structure. A university student is primarily accountable to the university for academic use; the same person may also be accountable to an employer for workplace use and to parents if still a minor. ### 2. What Each Layer Must Do Every parent authority is required to: - Publish a clear, accessible AI use policy appropriate to its context. - Ensure reasonable access for those under its care (equity of opportunity). - Provide age- or role-appropriate literacy and safety guidance. - Set disclosure and integrity rules where relevant. - Maintain basic oversight and a simple escalation path for harm or misuse. - Review the policy periodically against evidence of benefit and risk. Government’s role is to define the minimum standards each layer must meet and to designate the residual public authority for those without a private parent. ### 3. The Residual Case: People Without a Parent Authority Not everyone belongs to a family that can exercise meaningful oversight, an educational institution, or a formal organisation. For these individuals the residual duty falls to the **local public administration**. Practical mechanisms: - Local authorities maintain a lightweight public AI access and literacy service (physical or digital points, short mandatory orientation modules, basic safety guidance). - Registration or simple verification can be used where high-risk tools or high-volume access are involved, but mass surveillance or licensing of ordinary use is avoided. - The local office becomes the default “parent of last resort” for reporting serious misuse, supporting vulnerable users, and ensuring baseline access is not denied purely on the grounds of lacking institutional affiliation. - National framework legislation designates which tier of local government holds this residual duty and how it is funded and overseen. This prevents a permanent underclass of unregulated or completely excluded users while keeping the primary responsibility as close to the individual as possible. ### 4. Principles That Remain Constant Across All Layers - Access is the default; restriction requires justification. - Critical literacy and the right to decline use are protected. - Data protection and privacy standards apply uniformly. - Accountability is real: parent authorities can face regulatory or civil consequences for systemic failure to set and enforce reasonable rules. - No parent authority may outsource its duty entirely to a technology vendor. ### 5. Government’s Proper Role Government does not replace the parent layers. It: - Legislates the hierarchical duty structure and the residual local-authority obligation. - Sets the minimum content of policies (integrity, privacy, literacy, equity, review). - Empowers a national or regional oversight body to audit systemic failures and issue binding guidance. - Funds shared literacy resources and residual local services. - Avoids dictating specific tools or pedagogical/organisational methods. ### 6. Expected Effects - Clear lines of responsibility instead of diffuse “everyone and no one” accountability. - Institutions and organisations treat AI governance as part of their ordinary duty of care rather than an optional innovation project. - Families receive recognised standing and practical support rather than being left to improvise. - Individuals outside formal structures still have a defined public backstop. - Innovation and legitimate use remain possible because regulation stays principles-based and layered rather than centralised and prescriptive. **Summary statement for adoption** Generative AI access and use shall be governed by the nearest competent parent authority—family, educational institution, or organisation. Heads of those authorities are accountable for setting and upholding appropriate rules. Where no such authority exists, the designated local public administration assumes residual responsibility. National law establishes the framework, the minimum standards, and the residual duty; it does not micro-manage tools or everyday decisions. This structure scales from the household to the workplace to the university while closing the gap for those who would otherwise fall through.