The arrival of autonomous agents is evolving the very notion of sovereignty and infrastructures in the digital age. Faced with a rise in technology often faster than old rules, it becomes ess
The arrival of autonomous agents is evolving the very notion of sovereignty and infrastructures in the digital age. Faced with a rise in technology often faster than old rules, it becomes essential to give a legal existence to algorithmic entities. Estonia, a pioneer in innovative institutions and 100% digital, confirms once again as the global laboratory of this transition. By granting an official status to artificial intelligences, the Baltic state traces a new path in the history of technological governance.
On Wednesday, June 17, Kristen Michal, Estonian Prime Minister, officially approved a proposal from the advisory council concerning the establishment of an “AI personal identification code”. This historic measure will give artificial intelligence agents their own state digital identity, different from that of natural persons, companies, or institutions that own or operate them.
Through this approach, the head of government intends to place his country at the forefront of global technological regulation. Kristen Michal publicly stated on social network X that Estonia could become “the first country in the world to equip AI agents with an official digital identity”.
Operationally, this system aims to solve an ongoing structural problem: the current requirement for an algorithm to borrow the global digital identity of its human owner to accomplish routine tasks. The new system will allow machines limited, controllable, and auditable rights. Far from giving blind and permanent access to the personal data of their principal, this directive will limit the agent to strictly compartmentalized acts :
This transition towards identifiable agents occurs within a state ecosystem where artificial intelligence is already operationally integrated. Since the launch in January of its national AI development program, “Eesti.ai“, the Estonian government has been deploying autonomous systems in schools and ministerial services.
For this, the country notably relies on Bürokratt, a tool officially defined by the administration as an “AI-based digital assistant created by the state to help institutions provide modern and efficient customer service”. This massive deployment made the emergence of an internal IT access compartmentalization protocol necessary.
To explain why it is so important to legally regulate these interactions, Prime Minister Kristen Michal shared his vision of the technological future on social media: “In the future, artificial intelligence will perform digital actions on behalf of a person, a company, or an institution: drafting reports, preparing statements, or communicating with information systems. But we need to know who acts, on behalf of whom, with what powers, and who answers”. The requirement for such a framework is reflected internationally, where private initiatives like Sam Altman’s decentralized World network also try to structure the distinction between human and robotic requests.
Start your crypto adventure safely with BybitThis link uses an affiliate program.This high degree of bureaucratic integration relies on a mature and decentralized digital infrastructure. In December 2024, Estonia moved 100% of its public services online and has always relied on the KSI blockchain (Keyless Signature Infrastructure) to ensure the integrity of its judicial, health, and property registers since 2012.
The country also chains world firsts, having declared the internet a universal service as early as 2000, paving the way for legislative elections in 2023 where most votes were cast online. This immutable and fully digital architecture will serve as the technical foundation to host and transparently audit the signatures and transactions of AI identities.
However, the absence of clear safeguards on financial responsibility can have serious consequences in these connected networks. The latest case involved an unsupervised autonomous agent that created an unexpected $6,531 invoice on Amazon Web Services (AWS) in less than a day, after scanning a network without human validation, forcing its creator to launch a crypto donation campaign such as bitcoin to cover the damage. This drift perfectly exemplifies the risks of software autonomy when interacting directly with financial or cloud infrastructures without any precise legal accountability framework.
Estonia’s advancement opens major perspectives as well as complex questions about the future of digital law. By concretizing the identity of algorithmic agents, Estonia proposes a model likely to significantly reduce risks of misconduct and fraud, while securing economic interactions within the framework of Web3 and the AI economy. However, a nuanced analysis requires highlighting the project’s dark areas. No official implementation date has been announced, and the government has not provided any details on the distribution of financial or legal responsibility in the event of costly errors committed by an AI with its own code.
