Article 12 and digital governance
Article 12 and AI Systems
When a public function is carried out through an AI model, software workflow or contracted technology, constitutional responsibility should remain traceable. This page treats that position as constitutional analysis, not as a claim that every question is settled law.
The question
Article 12 matters because it helps locate the authority answerable for public power. If an AI system influences recognition, service access, risk classification, benefit delivery or grievance handling, the citizen still needs an accountable public authority, a record of the decision and a route to remedy.
What the record supports
- The existing Article 12 page frames public authority as the first question of accountability.
- The Digital Constitutional Personhood framework requires explanation, correction, human review and traceability when digital systems affect rights.
- The intelligence archive contains case files where portals, records, certifications and grievance systems are tested against citizen outcomes.
Limits of the present evidence
This page does not identify a specific deployed AI model or assert unlawful automated decision-making by a named institution. It establishes the archive's accountability test for future AI-related records.
How AI systems enter public functions
AI systems enter public functions in several ways. A ministry or department may procure a model from a private technology vendor and integrate it into a grievance portal, a benefit eligibility system, a risk-scoring tool or an identity-verification process. Alternatively, a contracted operator may deploy AI within a larger digital infrastructure that the government agency nominally controls but does not technically manage. In both cases, the constitutional question is the same: has the public function — the exercise of power that affects a citizen's recognition, service, liberty or remedy — been delegated to a system that remains invisible to the citizen and unaccountable to any named public authority?
Article 12 was designed for a world in which governmental power was exercised by identifiable officials, institutions and statutory bodies. The DISHA research position is that its logic extends naturally to AI-mediated governance: the citizen's constitutional relationship is with the State, not with the technology the State chooses to use. If a state-controlled portal uses an AI model to rank grievances, deny applications or flag identities, that AI model is operating within the constitutional perimeter of Article 12, and its outputs must meet the same accountability standards as any other public administrative act.
The contractor and vendor question
One of the recurring complications in AI and public governance is the role of private technology contractors. When a company develops and operates an AI system used to deliver a public function, the citizen may believe they are dealing with a government service while the government may assert that responsibility for any error lies with the vendor. Article 12 analysis does not resolve this ambiguity in favour of either party's convenience. The archive's accountability test asks which entity controls the public function — not which entity wrote the code — and treats that controlling entity as the Article 12 authority responsible for explanation, correction and remedy.
The intelligence archive's AI accountability test
When the DISHA intelligence archive examines an AI-related public governance record, it will apply a consistent four-part test. First: can the affected citizen identify which authority is responsible for the AI system's output? Second: has an explanation of the system's logic or criteria been provided in terms a citizen can understand and challenge? Third: does a human review path exist that can override the system's output when it produces an incorrect or unjust result? Fourth: has the record of the AI system's decision been preserved in a form that allows later audit, correction or appeal? A case file that cannot satisfy all four questions will be recorded in the archive as containing unresolved accountability gaps under the DISHA claim-to-source system, pending public authority response.
Connection to Digital Constitutional Personhood and DISHA
The Digital Constitutional Personhood framework asserts that a citizen's digital identity — the data profile, biometric record, application history, grievance log and identity verification trail held by public systems — is an extension of their constitutional personhood under Articles 14, 19 and 21. An AI system that operates on that digital identity without accountability is therefore not a neutral technical tool; it is an actor within a constitutional relationship. The DISHA architecture's No-First-Use doctrine applies here: intelligence about AI systems in public governance is gathered to defend citizen rights, isolate accountability failures, preserve evidence and support recovery — never to exploit vulnerabilities or retaliate against institutions.