A panel of experts led by Yoshua Bengio and commissioned by the UK government has highlighted concerns about job losses, environmental impact, and regulatory challenges posed by advanced AI technologies. The report warns of significant short-term unemployment increases due to automation, expanding carbon emissions from datacentres, and potential regulatory lag in response to rapid technological advancements.
An international panel of experts, inspired by the Intergovernmental Panel on Climate Change, has released its inaugural report on the safety of advanced artificial intelligence (AI). The panel, chaired by leading computer scientist Yoshua Bengio and commissioned by the UK government, raised concerns about the environmental impact, potential job losses, and regulatory challenges posed by AI technologies.
The report highlights that advancements in AI, including tools like ChatGPT, could lead to significant short-term increases in unemployment by automating various tasks. However, some economists believe the job losses might be offset by new roles in non-automated sectors. An International Monetary Fund (IMF) paper cited in the report suggests that 60% of jobs in advanced economies are exposed to AI, leading to diverse outcomes ranging from job automation to job complementation.
Environmental concerns also surfaced, with the report noting the increasing carbon emissions associated with the growing use of datacentres needed to train and operate AI models. The report warns that AI could soon become the largest consumer of datacentre electricity.
Furthermore, the panel expressed apprehensions about the pace of regulatory responses potentially lagging behind rapid technological advancements. Other risks include data biases within AI models, predominantly reflecting Western cultures, and fears of losing control over AI systems.
The report, released ahead of next week’s AI summit in Seoul, underscores the uncertainty around the future of general-purpose AI, suggesting that both very positive and very negative outcomes are possible, and that societal and governmental decisions will critically influence the trajectory of AI technologies.
