How generative artificial intelligence impacts the climate and environment, and could contribute to the green transition, are current topics that the Norwegian Board of Technology wishes to explore.
Applications based on generative artificial intelligence can, through simple prompts, create new text, sound, and images. This powerful technology had its breakthrough in 2022, with the chatbot ChatGPT and the large language model (LLM) GPT-3. LLMs are machine learning models able to interpret and generate text, trained on huge amounts of text via deep learning algorithms. Both training and using the models require large amounts of computing power and energy, which leads to greenhouse gas emissions.
Artificial intelligence can help solve climate challenges
Classical machine learning excels in processing vast amounts of data and make predictions, a capability that has contributed to improved climate models. With generative artificial intelligence capacity to generate new data, these technologies can be combined to create more powerful and advanced tools. For instance, climate research can be accelerated by connecting disciplines, creating new knowledge, and making it available in new ways. Moreover, the technology can also enhance energy efficiency, as demonstrated by Google’s achievement of a 40 % reduction in energy consumption from data centers.
… but affects the climate and environment negatively
Generative artificial intelligence depends on the global Information and Communication Technology (ICT) industry, already a significant contributor to emissions and energy consumption. ICT equipment, data centers and digital networks collectively account for approximately 10 % of the world’s energy consumption. LLMs rely on specialized computing chips or Graphics Processing Units (GPUs) designed for high-performance computing. These chips contain critical minerals that are associated with both environmental degradation due to mining and to human rights violations. Besides being resource- and energy-intensive during production, the chips also play a role in the rivalry between the US and China.
Widespread adoption of generative artificial intelligence can result in heightened climate and environmental impact, requiring measures to make LLMs greener. On the other hand, this technology also holds potential to contribute significantly to the green transition. For instance, President Biden’s voluntary guidelines, require companies to develop AI systems aimed at addressing major societal challenges such as climate change.
In this project, we will consider the following questions:
- How can generative artificial intelligence affect climate and the environment?
- Can the technology be used to accelerate the green transition in Norway?
- Should and can Norway support the development of green generative artificial intelligence?
- What regulatory, political, and technical solutions are needed?
There are many unsolved questions related to the development and use of generative artificial intelligence. This project is part of a larger work with generative artificial intelligence. You can read more about this project here.