Today Roomba, tomorrow Optimus

The acceleration of AI is presenting us with an unprecedented transformation: the arrival of humanoid robots. What for decades was the stuff of science fiction is now beginning to take shape in real life. AI is now an omnipresent engine, powering everything from voice assistants to the creation of content on social media. But the real turning point lies in machine learning systems – the ability to train models with large volumes of data to learn, adapt and generalise behaviours. This evolution is allowing machines that once carried out rigid and predictable instructions to now act with a certain amount of interpretation and, to a certain extent, decision-making.

In our test of top-of-the-line robot vacuum cleaners this year, we saw how AI has already transformed a simple household appliance into an autonomous agent, capable of recognizing obstacles and learning routines. Although modest, this application illustrates what is at stake: how machine learning can provide machines with functional intelligence that goes beyond simply following commands.