This Monday, the first UN Global Dialogue on Artificial Intelligence Governance opened in Geneva, the first intergovernmental forum in history devoted to this issue, with all 193 member states at the table alongside the private sector, academia and civil society. Secretary-General António Guterres called for governing the technological transformation “before being governed by it” and proposed a global child protection pact: systems verified as safe for minors before their launch, zero tolerance for AI-generated child sexual abuse, and referral to human help when there are signs of distress.
In the plenary session, the Independent International Scientific Panel on AI, co-chaired by Yoshua Bengio and Nobel Peace Prize laureate Maria Ressa, delivered its central warning: current science cannot guarantee that the growth in model capabilities will not end up causing “catastrophic harm.” For Latin America the meeting matters twice over. As a milestone, because it is the first time AI governance is discussed in a space with universal participation. And as a mirror of an asymmetry: the region has a single formal leadership seat (El Salvador co-chairs the dialogue together with Estonia) and a single direct scientific voice, that of Chilean Loreto Bravo, of the Universidad del Desarrollo, one of three Latin Americans among the panel’s 40 experts. Bravo warned that AI’s productivity gains “are not automatic” without infrastructure and institutional capacity. Hours before the opening, no government in the region had confirmed a ministerial delegation of its own, unlike India, which sent one headed by its Minister of State for External Affairs.
Also today
- Alibaba bans Claude Code for its employees and Anthropic reports the circumvention of restrictions to the US Senate — Alibaba is ordering a migration to its own tool, Qoder, starting July 10, while Anthropic rolls out identity verification to block access through subsidiaries and VPNs. The fragmentation of the AI stack between blocs keeps escalating.
- Argentina debates a bill on AI-run “Automated Companies” — The initiative would create an unprecedented legal form for companies run by AI; specialists are divided between a necessary regulatory advance and a “false debate” that does not settle who is accountable when there is a human behind it.
- Mexico heads into the 2027 elections with no framework for bots and deepfakes — 67 of 85 AI regulation initiatives have been stalled in Congress since 2023.
- Samsung Foundry cements its role as a custom AI chip maker for Anthropic and Meta — It is evaluating a 2-nanometer process for an Anthropic chip of its own and manufactures the third generation of Meta’s accelerator, one more sign that the big labs are seeking to reduce their dependence on Nvidia.
In the region
The Geneva dialogue brings into view two paths the region is traveling at once. In multilateral forums, Latin American representation is thin: one country co-chairing and three experts on the scientific panel. In domestic regulation, Argentina and Mexico mark the two ends of the spectrum. Argentina is moving forward, amid controversy, toward a legal form for AI-run companies that almost no country has; Mexico, by contrast, documents its own paralysis with hard data (two thirds of its bills stalled) just as it faces midterm elections exposed to bots and synthetic content. It is the same question seen from two angles: where a country’s capacity to govern the AI already operating within its borders is really defined.
Threads we’re following
This adds to a story we have been following. A few days ago, the same UN scientific panel put numbers on the region’s absence: about 90% of the compute that trains the most advanced models is concentrated in just two countries, and no Latin American country appears on that map. The Global Dialogue that opened today is the space where that report begins to turn into political decisions. This week’s picture (countries with ministerial delegations on one side, a region with a minimal voice on the other) is the next chapter of the same story: whoever builds the infrastructure tends to be whoever sets the rules.
If global AI governance is decided with Latin America on the margins of the room, is the region better off building a collective voice in forums like this one, or is the ground where its technological sovereignty is at stake the domestic regulation that Argentina and Mexico are trying out today in such different states?
About this entry. It is generated automatically from public sources, without human review before publication. It may contain errors of interpretation or summary; please check each story against its original source (the links lead there) before citing it or making decisions based on it.
Doble Click is written with Anthropic models.