Colombia’s watchdogs launch AI transparency training for agencies

Colombia’s oversight bodies are moving early on a problem many larger jurisdictions are still theorizing about: how to make opaque algorithmic systems used in government legible and contestable to citizens. The joint directive and training program focus on transparency around how automated systems and AI influence public‑sector decisions, from what data they use to how outputs affect people’s rights. That is less flashy than frontier model benchmarks, but it is exactly the kind of institutional plumbing that will determine whether societies can live with increasingly powerful AI systems embedded in public services.([procuraduria.gov.co](https://www.procuraduria.gov.co/Pages/procuraduria-y-defensoria-fortalecen-transparencia-algoritmica-para-proteger-derechos-ciudadania.aspx))

For the race to AGI, this is a reminder that governance capacity is not the exclusive domain of the US and EU. Latin American regulators are experimenting with pragmatic tools to demand documentation, explanation and accountability from any AI used in public functions. If those tools work, they could be copied elsewhere and reduce the risk that black‑box systems silently shape welfare decisions, policing or benefits eligibility.

The initiative also shows civil‑rights institutions asserting a voice before AGI‑level systems arrive, rather than waiting to clean up after harms. That proactive stance may slow the most aggressive deployments in Colombian public agencies, but it should make the remaining ones more socially durable. In the long run, that kind of legitimacy will be crucial if states want to keep scaling their reliance on AI in core governance functions.