.png)
November 27, 2024
The use of generative AI tools like ChatGPT, Microsoft CoPilot, and Google Gemini is rapidly increasing. A recent study by CloudNine PR shows that 83% of the British population are familiar with these technologies. At the same time, awareness of the ecological impact of these advancements is growing. More and more people are demanding transparency from companies regarding energy consumption and the associated environmental costs.
Behind every AI-generated email, every image, and every chatbot dialogue are huge data centers with thousands of energy-hungry servers. These servers are responsible for both training the large language models that power generative AI and for processing individual user requests. Unlike a simple Google search, which requires comparatively little energy, a single query to a generative AI can consume up to ten times as much electricity.
The numbers are impressive: If all nine billion daily Google searches worldwide were replaced by generative AI queries, the additional electricity demand would equal the annual consumption of 1.5 million EU citizens. According to Morgan Stanley, the energy demand of generative AI will increase by 70% annually until 2027. At that point, the energy demand for generative AI systems could equal the electricity consumption of an entire country – for example, Spain, based on 2022 consumption.
The CloudNine PR study also highlights the growing awareness among British consumers of the environmental impact of generative AI. Almost one in five respondents said they did not trust AI providers to responsibly manage their environmental impact. Among regular users of these tools, 10% expressed a willingness to pay a premium for products or services that prioritize energy efficiency and sustainability. Interestingly, more than a third (35%) of respondents believe that generative AI tools should actively remind users of their environmental impact. This could lead to a more conscious use of the technology and increase pressure on companies to use more environmentally friendly technologies.
Fortunately, some companies and policymakers are already responding to these concerns. In the US, the "Artificial Intelligence Environmental Impacts Act" was introduced earlier this year. This law aims to standardize the measurement and reporting of CO2 emissions by AI companies. It also provides developers with a voluntary framework for assessing and disclosing the environmental impact of their systems, thus promoting more transparency in the industry.
Large tech companies are also showing initiative. Salesforce, for example, supports legislation that mandates standardized methods for measuring and reporting the carbon footprint of AI. Experts see various ways to reduce the environmental impact of generative AI: energy-efficient hardware, sustainable cooling methods in data centers, and the switch to renewable energies.
Despite these efforts, the need to address the environmental impact of generative AI remains urgent. Uday Radia, owner of CloudNine PR, puts it succinctly: "Generative AI has enormous potential to improve our lives, but it's a race against time to make it more sustainable before it gets out of control."
The rapid development of generative AI raises important questions regarding its sustainability. While the technology offers enormous opportunities to optimize processes and drive innovation, the associated energy consumption and resource requirements must not be ignored. It is essential that companies, policymakers, and society work together on solutions to minimize the environmental impact of this technology and ensure a sustainable future. The increasing awareness of consumers about this issue is an important step in the right direction. Only through transparency, responsible action, and the development of more environmentally friendly technologies can the full potential of generative AI be realized without jeopardizing the resources of our planet.
Bibliographie: - https://pressreleases.responsesource.com/news/105761/new-research-generative-ai-awareness-soars-to-among-brits-but/ - https://www.nature.com/articles/d41586-024-00478-x - https://e360.yale.edu/features/artificial-intelligence-climate-energy-emissions - https://www.rtinsights.com/generative-ai-adoption-soars-insights-from-mckinseys-latest-survey/ - https://curious.earth/blog/regenerative-ai/ - https://www.polytechnique-insights.com/en/columns/energy/generative-ai-energy-consumption-soars/ - https://www.pwc.de/en/digitale-transformation/generative-ai-artificial-intelligence/the-genai-building-blocks/generative-ai-unleashed-for-sustainability.html - https://www.economist.com/business/2024/04/11/generative-ai-has-a-clean-energy-problem - https://www.mightybytes.com/blog/is-ethical-and-sustainable-ai-possible/ ```