International. In recent years, Google has been testing an algorithm that learns how to adjust cooling systems (fans, ventilation, and other equipment) to reduce energy consumption. This system previously made recommendations to data center managers, who would decide whether to implement them or not, leading to energy savings of about 40 percent on those cooling systems.
Now, Google says, it has effectively given control to the algorithm, which manages cooling in several of its data centers on its own.
"It's the first time an autonomous industrial control system will be implemented at this scale, to the best of our knowledge," says Mustafa Suleyman, head of applied AI at DeepMind, the London-based artificial intelligence company it acquired in 2014.
The project demonstrates the potential of artificial intelligence to manage infrastructure and shows how advanced AI systems can work in collaboration with humans. Although the algorithm runs independently, one person manages it and can intervene if they seem to be doing something too risky.
Kava says managers relied on the previous system and had few concerns about delegating more control to an AI. Still, the new system has safety controls in place to prevent it from doing anything that has an adverse effect on cooling. A data center manager can observe the system in action, see what the algorithm's level of confidence is regarding the changes it wants to make, and intervene if it appears to be doing something improper.
Source: MIT Technology Review.
