Data center cooling is a critical part of the process, as data centers must remain cool to function properly. It is therefore essential that the cooling system operates continuously and is regularly inspected and maintained. In most cases, this work is still done by human hands, but if Google has its way, that won’t be necessary anymore. The company is working on an Artificial Intelligence system that adapts to a data center’s cooling system!
So what is an AI system like that doing in a data center?
The AI system used by Google essentially controls the entire process within the data center. It has learned to maintain the temperature within the data center as optimally as possible while also saving energy. Such an AI system saves an average of 40% in energy throughout the entire process, which is, of course, a significant gain. This is now an automated process, but it took Google a lot of work to get there. They reviewed the recommendations the system made and then applied them manually. This is how Google operated for several years before automation. The algorithm indicates when the servers should be cooled, ensuring that no unnecessary energy is consumed at times when it isn’t needed.
Even fewer jobs for people?
You might immediately think, “Great, even less work for humans,” but that’s not really the case here. There are always people on site to monitor the system and intervene when necessary. Every 5 minutes, the system takes a snapshot of the cooling system and sends this information to the neural networks. The AI then decides how to maintain the current temperature while minimizing energy consumption. However, these decisions are only implemented by on-site staff after a safety check. So people only intervene if the system wants to do something that could pose a risk. Humans don’t necessarily have to do much in this process, but someone must always be present to monitor the AI-controlled cooling within data centers!
Data center energy consumption must be reduced
This is a debate that has been going on for at least a few years: data centers consume too much energy. This was a major point of criticism, especially during the energy crisis, when prices skyrocketed. Fortunately, more and more companies are working on solutions to this problem, such as Google with its AI system. There are also pilot projects underway in the Netherlands to use the waste heat from data centers to heat homes. Google itself states that its AI system has already reduced the energy consumption of its data centers by about 40%. This is also a key priority for Google, as the company faced significant criticism last year regarding the energy and water consumption of its data centers in maintaining its AI applications. Consider, for example, new AI features in the search engine’s results. So the problem they’re trying to solve with AI is actually caused, in the first place, by their use of AI.
Other uses for Google's AI system
As the AI system continues to evolve, the tech giant expects it to be used for other purposes in the future. Consider, for example, combating climate change through advanced computational models or saving energy in other sectors. Particularly in industrial settings, where many processes are already automated, significant gains can be achieved through the use of AI.
Do you think AI technology can help us consume less energy and better manage the climate in the future? Or will this technology exacerbate the problem because AI itself also consumes a lot of energy?