Artificial Intelligence: Curse or Blessing for Sustainability?
The sustainability killers
The race for ever more powerful Artificial Intelligence supported search engines has a downside: it is accompanied by a drastic increase in demand for computing power and storage space. This in turn leads to increased energy consumption, which – if it does not come from sustainable sources – drives up CO2 emissions. The technology giants OpenAI and Google have not yet provided any information on the computing costs of their products. According to an estimate by researchers at the University of Berkeley, the training of ChatGPT alone generates more than 550 tons of CO2 equivalents – as much as a person flying back and forth between New York and San Francisco 550 times.
And training is not the end of the story: Martin Bouchard, co-founder of Canadian data center company QScale, estimates that his reading of Microsoft and Google’s plans for search will require “at least four to five times the computing power per search” when generative AI is added to the process. He points out that ChatGPT is currently shutting down its understanding of the world at the end of 2021 in order to reduce computing power.
The question is not whether, but how
According to the International Energy Agency, data centers are already responsible for around one percent of global greenhouse gas emissions – and the trend is rising. “Electricity consumption will increase, no ifs, ands or buts. The only question is how we will meet this additional demand: from ‘clean sources’ or from fossil fuels,” says Martin Schwab, CEO of CKW and alumnus of the Rochester-Bern EMBA.
Numerous promises have already been made in this regard: Microsoft has pledged to become carbon neutral by 2030. Google has also promised to achieve net-zero emissions across its entire operations and value chain by 2030. Martin Schwab: “The key to sustainable energy production globally is the massive expansion of renewable energies in the form of solar and wind power as well as nuclear energy as a backup. If we are prepared to use nuclear power, we can cover the additional energy demand without fossil fuels.”
Turning the tables: More sustainability through AI
At the same time, AI also offers opportunities for sustainability. It can recognize patterns, analyse large amounts of data and coordinate systems, which can contribute to greater efficiency and savings, especially in the power supply sector. The keyword here is smart grids – intelligent power grids that are centrally controlled and optimally coordinated to compensate for power fluctuations in the grid.
Electricity is difficult to store. What is produced has to be consumed at the same moment. One of the biggest challenges is therefore to keep the frequency in the grid stable. “It’s like a bathtub with millions of small holes that open and close. At the same time, there are tens of thousands of inflows. The task of the electricity producers is to keep the water stable,” explains Schwab.
AI can help to ensure stability, and the idea could be taken even further if it is extended to private consumers: “If households allow it, we could use AI to set up a nationwide system that is operated in line with the grid. This would mean, for example, that households could indicate that they don’t care when exactly the dishwasher runs in the next 24 hours, and the grid operator could start it when there is a lot of electricity in the grid,” says Schwab. AI has great potential here to make a complex system as efficient as possible.
We decide
New technologies are neither good nor bad in themselves. It’s what we do with them that counts. “When I graduated from the Rochester-Bern EMBA in 2001, the Internet was a relatively new and very big topic and we were discussing the possibilities it opened up for us,” says Schwab. Today, 23 years later, the Internet has become an integral part of our everyday lives and we talk about Metaverse, ChatGPT, Bitcoin etc. The question is always: what do we make of it?
