The way in which AI systems and services are currently designed, built and used places ‘demands on resources such as energy, water and critical materials, which can create new environmental harms or exacerbate existing harms’, said a report. — AFP Relaxnews
Governments need to start acting over the environmental risks, including potential water shortages, posed by the development of artificial intelligence, a UK body advising on engineering policy warned on Feb 7.
The National Engineering Policy Centre – comprising 42 institutions headed by the Royal Academy of Engineering – highlighted potential fallout from data centres relying heavily on energy, water and critical materials like silicon.
"These rapidly growing resource demands could have far-reaching effects, such as creating competition for renewable energy or drinking water," the NEPC said in a report that urged short-term measures by world governments to help provide sustainable AI.
These include setting environmental sustainability requirements for data centres, which require significant amounts of water for cooling purposes.
"Artificial intelligence can be leveraged to accelerate progress towards net zero carbon emissions and improvements in environmental sustainability," the report noted.
However, the way in which AI systems and services are currently designed, built and used places "demands on resources such as energy, water and critical materials, which can create new environmental harms or exacerbate existing harms", it added.
The report – entitled Engineering Responsible AI: foundations for environmentally sustainable AI – comes after British Prime Minister Keir Starmer last month outlined a vision to put the UK at the forefront of artificial intelligence development.
Britain presently has the world's third-largest AI industry after the United States and China.
In a warning shot to the UK and other countries, the chair of the report's working group said "to build systems and services that effectively use resources, we first need to effectively monitor their environmental cost".
Professor Tom Rodden added that "once we have access to trustworthy data pertaining to their environmental impacts... we can begin to effectively target efficiency in development, deployment, and use – and plan a sustainable AI future". – AFP Relaxnews
