AI agents may not be real people, but they now have an identity in Estonia. The small European nation is the first country to issue AI agents ID codes for identification. The digital ID codes will allow AI agents to act on behalf of people, organizations or companies but only within strictly confined limits and in ways that are accountable and auditable. Estonia’s AI ID initiative is precedent-setting and is likely to influence AI policies elsewhere.
AI is increasingly carrying out digital tasks on our behalf, notes Estonian Prime Minister Kristen Michal. “To that end, it must be clear who is acting on whose behalf with what rights and who is ultimately responsible.”
Michal says AI ID codes are needed to prevent situations where individuals or organizations are required to grant AI assistants access to all their data and AIs are given broad authority to act. Estonia’s initiative comes as research suggests traditional identity checks like multi-factor identification are unable to govern AI agents that can “act, decide and transact at machine speeds.” Digital systems must be able to distinguish between human and AI in order to establish digital trust.
Estonia may be a small country of just 1.3 million people but its involvement and influence in digital technology is outsized. Almost every aspect of public discourse with the government can be carried out online. AI has been a major element in public sector digitalization with over 130 projects involving the technology.
Estonia’s new AI ID program makes it possible to specify the degree to which an AI can act on a human’s behalf. For example, AI agents may be authorized to view data only, prepare a document, make a payment or act solely with specified financial limits. AI ID also is meant to forestall incidents where an AI wipes out a person’s data without approval.
Perhaps most importantly over the long term is that AI ID is designed to be a cultural preserver. Basically, Estonia wants AI to act like an Estonian. AI ID advocates like the Boston Consulting Group also argue that cultural preservation makes smart business sense lest AI reduce company branding and identity to a universal and indistinguishable sameness. Culture must be embedded into how AI systems operate, make decisions and interact with people. The goal is to avoid AI-induced mediocrity and blandness.
“Without cultural grounding, AI agent interactions fail to reflect the behaviors and experiences that differentiate companies and their brands,” says BCG. “Customer interactions become consistent but not distinctive, decisions become efficient but not principled, experiences become scalable but not meaningful. The system works. But it no longer reflects who the organization is.”
Estonia’s initiative also comes as AI agents test liability concerns. Advocates say AI agents must face financial and criminal penalties the same way humans do for fraudulent or harmful behavior. AI companies generally think AI should face no consequences for AI harms but recent legal cases involving AI culpability suggest the courts think otherwise. Air Canada, for example, was recently held liable for bad chatbot advice.
For Estonia, the power of AI means an identifier has to be more than just a line of code. AI must now have the equivalent of an identity card. They’re checking IDs.
