Contenuti
Multimediali
Imagine opening your fridge, thinking about what you could cook with what’s inside, improvising if something is missing, and adjusting the recipe as you go.
For you, that’s effortless.
Or imagine assembling a piece of furniture, noticing that a screw is crooked, and deciding to ask for a steel washer from your neighbour to fix it.
For you, that’s natural — but not for an AI.
AlphaFold, the artificial intelligence system that helped its creators win the Nobel Prize in Chemistry, can predict the structure of proteins better than any human. But it excels only at the task it was trained for — it is not versatile.
Current AI systems have not yet reached the level of AGI — Artificial General Intelligence: a term introduced in the early 2000s to describe a machine capable of matching an adult human in cognitive flexibility and competence.
We don’t know if — or when — we will build one. But tests already exist to measure how close a specific AI system is to that goal. Some of these tests are in front of you — and as you’ll see, they are not so different from the ones designed for us.