As Artificial Intelligence is being applied in all fields of our activity, it is not an exception that it is also applied to one with such a future as nuclear fusion.
The "theory" of nuclear fusion has been known for almost a hundred years, but putting it into practice is something very difficult that until recently could not even be attempted.
The enormous technological development of recent years allows us to see the possibility that in a not too distant day it can be achieved, and the most optimistic hope to achieve it between 10 and 20 years. Although that remains to be seen.
What seems to be in no doubt is that Artificial Intelligence can help control the complex processes that must be produced in order to generate energy through the fusion of hydrogen isotopes.
The people of Deep Mind (it is the most advanced company in the world in Artificial Intelligence technology) is collaborating with researchers working on nuclear fusion processes to try to be able to control them and obtain energy.
One of the main problems is being able to control the fusion plasma inside the reactor. The magnetic coils must be controlled so that the plasma at temperatures between 150 and 300 million degrees does not collide with the walls, and this must be done thousands of times per second.
As nuclear fusion is also an unknown physical process that varies over time, the AI system has to learn by itself to achieve results, and that the energy obtained is maintained over time. And machine learning is one of the strengths of AI algorithms, where Deep Mind has already succeeded in the past.
It remains to be seen what is capable of being achieved, because so far all the fusion tests carried out have lasted a few seconds, and almost none have generated positive energy.
There is much work to be done, but it seems there are reasons to be fairly optimistic. Hopefully these expectations are met and we are able to achieve clean energy. That will change our life
EN 
ES_ES 