“Without Haste, but Without Pause”: How Uruguay’s Parliament Is Approaching AI
A conversation with Rodrigo Goñi Reyes, Speaker of Uruguay’s Chamber of Deputies, at the 152nd Assembly of the Inter-Parliamentary Union in Turkey
BY BEATRIZ REY
When Rodrigo Goñi Reyes, Speaker of Uruguay’s Chamber of Deputies, sat down with me on the sidelines of the 152nd Assembly of the Inter-Parliamentary Union (IPU), the conversation quickly moved beyond the familiar distinction between AI adoption and regulation. What stands out in Uruguay’s approach is not that it is pursuing both, but how it is sequencing and connecting them through a concept Reyes returns to repeatedly: the “enabling condition.”
For Reyes, the question is not simply how to regulate AI across society and within parliament, but how to structure regulation so that it actively supports its use. “We always talk about regulating as an enabling condition, that is, creating the conditions so that everyone can use artificial intelligence in the best way.” This shifts the role of regulation from a reactive exercise (focused on risks and constraints) to a forward-looking one that builds the institutional, technical, and economic conditions under which AI can be meaningfully adopted.
This framing has direct implications for how parliaments organize their own work. It begins with a reversal of a common expectation: rather than waiting for regulatory clarity, legislative institutions are expected to move first. As Reyes puts it, “a parliament cannot demand effort from society to adopt AI if it does not adopt it itself.” In Uruguay, this has translated into an ongoing process of internal adoption, including “legislative drafting, all transcription, and processing of what each legislator and each parliamentary debate produces,” alongside efforts to build capacity in emerging areas such as “prediction” and “impact evaluation.”
At the same time, the enabling condition strategy resists the push toward immediate, comprehensive regulation. Uruguay has already initiated a broad legal process — mandating input from public and private actors across domains “from copyright to child protection” — but Reyes is explicit that this does not translate into urgency for a single, overarching law. “We are not in a hurry to create a general regulation,” he notes. Instead, “the path gaining traction is responding to sectoral demands,” particularly in areas such as health, education, and security, where concrete use cases are already generating pressure for rules. Regulation, in this model, follows practice and evolves alongside it.
This approach is closely tied to concerns about institutional trust. The risks associated with AI in parliamentary settings are not hypothetical. “If…the Parliament has a hallucination event…the system collapses,” Reyes warns. This emphasis on risk containment shapes the pace and design of adoption, including the use of “closed systems” and controlled data environments. The result is a strategy that seeks to balance momentum with caution — what he describes as moving “without haste, but without pause.”
If enabling conditions define the domestic strategy, international cooperation emerges as a necessary extension of it. Throughout our exchange, Reyes returned to the role of the IPU and similar organizations as a coordinating platform for AI adoption across legislatures. “I truly believe that today more than ever, the IPU has a key role…because it holds the key,” he argues, pointing to its capacity to facilitate shared infrastructures and trusted collaboration.
The core issue is one of coordination. As I noted in our conversation, this is a collective action problem because the benefits of sharing data, tools, and lessons from early adopters are widely distributed, while the costs and risks of building and opening those systems are borne individually. Without coordination, each parliament has incentives to move cautiously or act alone, leading to duplication, fragmentation, and uneven capacity. Multilateral institutions are therefore necessary to ensure that no parliament lags behind and that all can learn from early adopters.
Reyes’ response is concrete: “all our data from each parliament…can be made available through the IPU, because we all benefit. Without that data, it cannot be done.” In this view, the IPU is not only a forum for exchange, but a potential repository of data, systems, and institutional trust necessary for scaling more advanced uses of AI, including real-time evaluation and predictive analysis.
This conversation also connects directly to a broader research agenda I am developing on the reimagination of politics under conditions of technological transformation. When asked whether politics can be reimagined today, Reyes is equivocal: “that answer…is not there.” What is clear, however, is that the underlying structure of political life has already shifted: “everything is overturned…the old order is gone. The new one is not yet formed.”
Rather than offering a definitive model, Reyes points to the emergence of new dilemmas that political institutions must confront. Central among them is the question of how to navigate the boundary between technological adoption and the preservation of human capacities. As he puts it, politics must now grapple with “what of the new technologies we have to adopt and what of the human we have to preserve.”
Beatriz: How is Uruguay’s Parliament approaching AI across both adoption and regulation, and what role does the idea of the “enabling condition” play in that strategy?
Rodrigo: Uruguay, like other countries, is moving along two lines with AI: adoption and regulation. Uruguay has always been pro-innovation, so we approach regulation as an enabling condition — that is, creating the conditions for everyone to use artificial intelligence in the best possible way, both internationally and in key areas such as entrepreneurship and business. Today, the main driver of productivity and competitiveness lies in new technologies. In the past, it was tariffs or traditional productivity. Now, the difference in productivity and competitiveness comes from the intensive and appropriate use of AI.
We approach regulation with that in mind: how to create the conditions for individuals and businesses to use AI effectively. The concept we use is central — the “enabling condition.” Uruguay, like other countries, developed an AI strategy and, based on it, a multi-stakeholder legal recommendation. At the beginning of 2024, parliament passed a law asking all public and private entities to provide input for this recommendation. It is very comprehensive, covering all areas of AI regulation, from copyright to child protection. But for now, our focus remains on the enabling condition.
I also insist that a parliament cannot ask society to adopt AI if it does not adopt it itself. In Latin America, we are somewhat behind, with the exception of Brazil and Chile. Following the Chilean model, the Uruguayan parliament is now in an intensive phase of adoption. This includes legislative drafting, as well as the transcription and processing of everything produced by legislators and parliamentary debates, so that citizens can easily access it. We are also working on prediction, using AI to generate rapid assessments for impact evaluation.
At the same time, adopting AI in a parliament involves risks, and precautions are necessary. If we are trying to build trust — encouraging widespread use of AI — and the parliament experiences a hallucination event, whether in information, citizen-facing outputs, or legislative drafting, the system collapses. That is why we are working with new systems from large companies that allow us to operate in closed environments. We believe in moving “without haste, but without pause.” The temptation to appear advanced without taking precautions is not worth it. Uruguay is moving forward firmly in modernizing these processes, and by this year we expect all of them to be operational. But we are proceeding with caution. On the regulatory side, Uruguay has approved a legal framework in which all legislators commit to a multi-stakeholder, consensus-based process.
You might say that one law can always be replaced by another. That is true, but Uruguay has chosen to approach this as a state policy. Within that multi-stakeholder logic, we are not in a hurry to adopt a general regulation. We have a digital caucus working steadily. Why? Because we do not see a clear benefit in rushing to pass a law. The approach that is gaining traction is to respond to sectoral demands (particularly in health, education, and security) where concrete needs are already emerging. These demands are already pushing for rules that allow progress, and that is where we are focusing. For example, the Ministry of Security is asking whether certain AI-based surveillance systems can be used. We are working to enable that. That is our general overview.
Beatriz: AI production for draft transcription is common, but for prediction and impact evaluation it is not. Can you explain how Uruguay is using AI for impact evaluation and prediction?
Rodrigo: These are predictive systems, essentially. The systems they use allow for the simultaneous evaluation of impacts. For example, the kind of real-time evaluation Australia is carrying out with platform ban (since 2025, Australia has adopted a ban on social media for those under 16) is what we aim to develop.
We will be able to assess social impact, and at the budget level, economic impact as well. But all of this predictive functionality is something we are trying to adopt with the precautions I mentioned, using our own data. This is also why we believe the IPU has a key role to play. With systems like Google Enterprise, what greater security could we have than coordination through the IPU, ensuring that each parliament works within trusted systems? That is one of the keys. I truly believe that today more than ever, the IPU has a central role in AI adoption in parliaments, because it holds the key.
And it holds the key because parliaments have always cooperated through the IPU. So what is the most direct way for me to access Australia’s real-time evaluation of its law? That law was passed in December 2025. It has only been four months (although it feels like ten years). I recently met with Australian counterparts, and they are continuously evaluating the policy using their own systems, and they are willing to share these.
At the same time, we need to ensure protection. It is not about producing general predictions — for example, what would happen if platforms were banned for minors under 16 — but about enabling each parliament to authorize access to these systems in a secure way. That is why we have a technology group meeting this afternoon at the IPU Assembly, and our proposal will move in that direction. I think there is real willingness, but sometimes things are not done simply because they are not proposed.
Beatriz: Your argument is that the IPU can serve as a repository of experiences and contacts.
Rodrigo: And systems. All our data from each parliament — I think the conditions are there for it to be shared through the IPU, because we all benefit. Without that data, it cannot be done. The question, then, is what platform allows us to build such a repository. I am convinced the answer to this question is possible and straightforward.
Beatriz: I think this is a collective action problem that can only be solved through multilateral and/or international institutions.
Rodrigo: And what the IPU has is trust, at a time when, as was said here and by the UN Secretary-General, many multilateral institutions are facing a loss of credibility. In the case of the IPU, which has always played a cooperative role in parliamentary work, the opposite is true. This creates a major opportunity — not one driven by novelty or glamour, but something truly decisive. If we want rapid AI adoption with the protections parliaments require, then this kind of shared repository is essential and the IPU can hold the key. I do not see any parliament refusing, because we all benefit quickly.
Beatriz: How do your colleagues in Uruguay view AI adoption? Are they all as enthusiastic as you? Do they understand it?
Rodrigo: As with everything, there is always a small group that leads. What we try to ensure is that this group exists. I strongly believe in digital caucuses. For example, in ParlAmericas, the Latin American regional parliament, we lead a digital caucus where we encourage all parliaments and all political parties to have an advanced group. This is necessary because parliaments are large, and many people naturally focus on day-to-day issues. But this topic requires specialization — it demands constant updating, and it is new. I want to get more involved with ParlAmericas now with the aim of leading its digital caucus. We really need to push this agenda forward; if we don’t approach it properly, democracy will be left behind on these issues.
Beatriz: You are also involved with Uruguay’s Committee of the Future. How important are these institutions?
Rodrigo: Through the Group of Latin America and Caribbean Countries (GRULAC) at the United Nations, we proposed that the IPU strengthen these spaces, because Committees of the Future apply anticipatory governance. It exists in Uruguay, Finland, and Chile. These committees allow not only rapid updating — they shape how you think. They focus on what is emerging and help identify new signals. They also incorporate a strong multi-stakeholder approach. The Committee of the Future in Uruguay has been operating for seven years, and it works in a multi-stakeholder way. The private sector is essential; not just as a participant, but as a necessity.
The IPU already has a technology group, but I think it lacks this anticipatory methodology. That is what future commissions provide. If you ask how to create them everywhere, the answer is simple: they need to be proposed. Every time I propose this, people agree with the idea. They recognize the need. And now this is no longer just a technical issue; it has become political. Technology appears in newspapers, on television, in daily news. That creates incentives for politicians to engage. But it still requires specialization, just like health or security. Here, the field is AI and technology. We need the right platforms, and I think the IPU has a major opportunity to respond.
Beatriz: You are the Speaker of the Chamber. How important is leadership sensitivity for adoption?
Rodrigo: For me as Speaker, it is the number one priority. A public opinion survey released yesterday shows that 80% of Uruguayans are very concerned about the risks of technological change. They expect Parliament to act at a time when trust in parliaments is low. That is a concrete demand. And if parliaments do not engage with this issue, they risk becoming irrelevant. I sometimes tell colleagues: if we do not address this, we will become a museum — a beautiful palace, but a museum. The conditions are there; what is needed is to create the spaces and the tools. That is what Committees of the Future are for.
Beatriz: What have you learned from Brazil and Chile?
Rodrigo: Brazil is unique. It has strong leadership, regardless of the government. It invests in this area and maintains technical continuity. Brazil is a global leader. Chile is also a leader, in a different way, particularly through its future commissions and regulatory processes. Other countries have the opportunity to learn from both. Chile is more open in that sense. Brazil, given its leadership, could do more to engage with the rest of Latin America. That would benefit both Brazil and the region.
Beatriz: I have a research agenda on reimagining politics. How would you reimagine politics today?
I think that answer — and this is something I think about constantly — does not exist. What is clear is that representation has changed substantially. What is clear is that everything is upside down. The classic traditional intermediation of representation has been shattered. Now, I think the new has not yet arrived.
Among the new things we need to incorporate into parliaments and politics, there is an area that was very far from politics: bioethics. Bioethics used to be a discipline focused on certain issues such as euthanasia, abortion, and assisted reproduction. But now bioethics is a discipline that necessarily has to help us, as politicians, to reimagine politics. For me, the big question is: What do we need to adopt from new technologies? Which elements from humanity do we need to preserve? That fundamental question has not been answered by any world leader, by any political leader, or by any parliament. The problem is not that this question has not been answered. The problem will arise if we do not confront it.
In the Uruguayan parliament, we created a National Bioethics Committee for these new issues, which are essentially the new technologies that also impact and modify the human condition. I will give you one example. A year ago, if you asked attendees in the IPU Assembly if we should ban social media, 90% would have told you: absolutely not because this hurts freedom of expression. Today, maybe 90% would tell you: something has to be done. We are seeing that the human condition — consciousness and critical thinking, which are conditions for freedom — is being irreversibly damaged by the constant presence of platforms. It’s good that evidence is now emerging that essential human characteristics — but also characteristics of democracy — are being irreversibly damaged, and that something must be done.
If you ask me, I do not dare to tell you how I would reimagine politics, but I can tell you that there are new questions, new dilemmas, and new areas of reflection that this new politics must incorporate. Before, we discussed whether we needed more or less state or how to achieve social justice. Those almost become secondary topics in the face of the enormous potential of new technologies. Politics must also discern which technology to adopt, how to adopt it, and what from humanity to preserve.
Modern Parliament (“ModParl”) is a newsletter from POPVOX Foundation that provides insights into the evolution of legislative institutions worldwide. Learn more and subscribe at modparl.substack.com.
