Servers used for mining cryptocurrency or powering AI consume a significant amount of electricity. In recent years, the rise of cryptocurrency has sparked concerns about its environmental impact. Mining cryptocurrency requires substantial energy expenditure. To generate a new bitcoin or other cryptocurrency, complex equations must be solved through a process called mining. Individuals and companies have scaled up their operations, buying high-powered computer chips called GPUs to mine cryptocurrency at a faster rate. This surge in demand for GPUs has also led to a chip shortage that affected various industries. The energy consumption of bitcoin mining alone surpasses that of entire countries like Norway and Ukraine combined. Although the environmental consequences of the crypto craze are still being assessed, the environmental impact of AI, which also heavily relies on GPUs, has received comparatively less attention. The AI revolution, particularly tools powered by GPUs, such as ChatGPT and Google Bard, demands a discussion on its environmental footprint, according to Sasha Luccioni, an ethical and sustainable AI researcher at Hugging Face. Quantifying the environmental impact of AI is challenging due to limited disclosure from companies about the energy use of their systems and the intangibility of AI itself.

While the detrimental effects of activities like driving a car are apparent, the impact of AI remains largely invisible to end-users. Significant amounts of water are used to cool data centers that house AI infrastructure, but this water consumption often goes unnoticed. Estimates suggest that training models like GPT-3 can consume millions of liters of water, and the energy consumption and resulting carbon footprint can be substantial. Efforts are being made to develop AI systems with reduced energy consumption, though there is a trade-off between performance and environmental impact. Without a radical reevaluation of AI development practices, the energy consumption of AI tools could exceed the energy use of the entire human workforce by 2025, according to technology analysts Gartner. By 2030, machine learning training and data storage could account for a significant portion of global electricity consumption. Treating AI with increased awareness of its environmental impact, similar to how we perceive cryptocurrency, is crucial. Balancing the awe-inspiring capabilities of AI with its environmental consequences is a necessary step forward. In other news, Elon Musk's company X (formerly known as Twitter) is facing legal issues, and the UK's Competition and Markets Authority may reconsider its decision to block a merger between Microsoft and Activision Blizzard. Moreover, a 1970s programming language called Prolog is being used to optimize strategies for winning the UK's national lottery, and smartphone etiquette guidelines are outlined by Paula Cocozza.