
Hong Kong ranks among the world’s largest data centre hubs but has a carbon footprint exceeding the global average, a United Nations think tank study has found, urging a responsible strategy to tackle the “unintended impacts” of using artificial intelligence (AI).
The June report by the United Nations University Institute for Water, Environment and Health also quantified not only carbon but water and land footprints of electricity arising from the use of AI.
The report is a call for using AI “responsibly and addressing its unintended impacts proactively to make it sustainable and equitable”, said Kaveh Madani, an environmental scientist who led the research team.
According to the report, every kilowatt-hour of electricity used to train an AI system also carries water and land footprints, in addition to carbon emissions.
From cooling heat-intensive data centres and generating power with water, to consuming land for energy infrastructure and fuel extraction, these processes could deplete water and land resources if not managed well.
Once an AI model is in use, continuous operations stemming from everyday user prompts account for 80 per cent of its total energy use, according to the report.