Argentina's government under President Javier Milei has launched a so-called social digital twin artificial intelligence system to reshape public policy. But the announcement video itself was marked by grammatical errors, clear signs of AI-generated content, and a deepfake of a minister.
Errors in the rollout
The video, meant to introduce the AI platform, contained multiple mistakes. Viewers spotted misspellings and awkward phrasing that looked like they came from a language model. More troubling was a clip that appeared to show a minister speaking — but it was a deepfake. The government hasn't explained how the deepfake was made or why it was used.
What the social digital twin is supposed to do
The system is described as a digital replica of Argentina's social and economic landscape. Officials say it will simulate policy outcomes before real-world implementation. The idea is to test changes on a virtual population — tweaking taxes, benefits, or regulations — and see what happens without affecting actual people. Similar tools exist in other countries, but Argentina's version is among the first to be built by a national government from scratch.
Unanswered questions about the technology
Few details have been released about the data powering the twin. Who built it? What algorithms does it run on? The errors in the announcement raise doubts about the care taken in developing the system. Critics inside Argentina, though not quoted directly, have pointed out that a tool meant to guide policy should be presented without basic mistakes.
The deepfake is a particular concern. If the government can't be transparent about a promotional video, trust in the AI's outputs may suffer. The minister whose likeness was used has not commented publicly. It's unclear whether the deepfake was an internal production error or a deliberate creative choice.
The project is still in its early stages. No timeline has been given for when the social digital twin will start producing policy recommendations. The government has not responded to requests for clarification about the video's production. Until those questions are answered, the system's credibility — and the policies it might suggest — remains an open question.
