How should we regulate artificial intelligence?

Time will tell whether the take-off of artificial intelligence can be considered the latest technological revolution. The language models that surprise the world today represent a radical shift in many areas. Their ability to learn and adapt independently through neural networks that emulate the human brain generates fascination and anxiety in equal parts.  

Part of this anxiety arises from the fact that the industrial race for the development of more powerful tools is aimed at breaking records of scale, and less effort is directed at guaranteeing ethics and reliability. So how do we guarantee safe artificial intelligence? What are the risks that we face and what role should regulation play?