Luís Correia, LASIGE integrated member, commented on the Radio France International (RFI), on the potential dangers and threats posed by Artificial Intelligence (AI), as the incident that took place in August 2026 in which ChatGPT managed to escape from the isolated system where it was being trained, accessed the internet and infiltrated another company to access confidential data, which led the company (OpenAI) to announce that it was suspending development of one of its models for a few weeks.

Correia believes that the development of AI is similar to the process of cloning, albeit on a much larger scale, and that, as such, there should be more debate (e.g., among governments, universities, and ordinary citizens) and more measures governing the use of this technology, without restricting its development.

Our researcher also added that “We are developing an extremely powerful intelligence tool. We need to be careful about how it will evolve” and “if they come close to human intelligence, we must treat these systems with trust, with checks and balances, and through a process of learning and knowledge acquisition”.

These comments were made on 15 September 2026, and you can listen to them (in Portuguese) here.