Our senior associate at az Tech, Antonia Nudman, spoke with Diario Financiero, where she analyzed the challenges facing the Chilean government in light of the widespread adoption of autonomous artificial intelligence agents.
The submission of more than 38,000 legal documents through an external technology provider to the Judiciary’s virtual office suddenly increased its workload. This development raised questions about the technical and human capacities of the country’s public institutions to cope with the widespread adoption of automation and autonomous artificial intelligence (AI) agents.
Experts and lawyers agreed that government agencies—to varying degrees—have gaps in technology, capabilities, and usage guidelines to address the new landscape imposed by AI. To address this, they proposed that institutions must adapt their systems, define common standards, and implement solutions to handle large-scale automation operations.
Tomás Vera, a professor in the Department of Computer Science at the University of Chile, stated that the main problem facing public institutions today is not the technology itself, but rather the differences among them. While some have advanced equipment, others rely on outdated computer systems, fragmented records, limited monitoring capabilities, and a “heavy dependence” on vendors.
“Institutions must prepare themselves because a growing proportion of interactions will be carried out by AI agents rather than people. This requires updating both the technological infrastructure and management processes,” he said.
Vera, who is also a scientific advisor to Congreso Futuro, noted that “capabilities are crucial, because a law can require security and accountability, but it cannot detect an overload, manage a backlog of documents, or restore a crashed system.”
Antonia Nudman, a senior associate at az Tech, said that some progress has already been made, such as the Guidelines for the Use of Artificial Intelligence Tools in the Public Sector, issued by the government in 2023, and the National AI Policy, but that progress must be made toward “more consolidated” cross-cutting rules and standards for the government.
According to the attorney, a “common baseline of capabilities” is required in the public sector—encompassing human resources, infrastructure, governance, cybersecurity, data protection, and the responsible use of AI—as well as cross-functional training for civil servants and multidisciplinary teams to address new technologies such as autonomous agents.
“It is not enough for some institutions to have advanced infrastructure or specialized equipment if others maintain more vulnerable systems or lack the capabilities to identify and manage new technological risks (…) The use of technology can no longer be confined solely to technical areas, because its risks and effects cut across the entire institution,” he stated.
Coordinated Actions
Vera said that a coordinated response from the government—and “not just isolated actions”—is needed to address the advancement of AI agents. To this end, he proposed that the National Cybersecurity Agency (ANCI) establish mandatory standards for resilience, monitoring, and response to large-scale automation, and that the Secretariat of Digital Government translate these into a common architecture for public platforms.
He added that this initiative should incorporate a “distinct identity for agents,” official application programming interfaces (APIs), usage limits, processing queues, elimination of duplication—repeated copies of information—traceability, and separation of critical services.
Regarding investment, the scholar proposed that the Budget Office (Dipres) and the Digital Government could prioritize technological modernization projects—such as monitoring centers, data management, and operational continuity—and recruit specialized talent in AI, while ChileCompra should raise the standards for security, auditing, load testing, and data control for government suppliers.
According to Vera, the Ministry of Science, Technology, Knowledge, and Innovation (CTCI) should continue to develop policies, risk assessment methodologies, and training in AI, but “the practical implementation should remain in the hands of each public agency.”
Meanwhile, Esteban Ruiz, executive director of the Chilean Legaltech Association (Altech), proposed that the Judiciary create special channels for automated submissions and bulk inquiries, complete with authentication and authorization mechanisms, as well as user registration procedures and access restrictions.
He also said that the Virtual Judicial Office “must evolve” toward an architecture capable of interacting with automated systems—such as AI agents—and that the institution should define a “comprehensive policy” for the use of AI and automation, as well as “clear rules for high-volume operations.”
Along these lines, Nudman noted that, in the case of the Judiciary, the Cybersecurity Framework Law requires it to adopt security measures for its networks and computer systems and authorizes the Supreme Court to issue regulations for this purpose.
The attorney added that, based on “common minimum standards,” each public institution should develop specific rules according to its functions, the technologies it uses, and the risks it faces.
