AI Summary
Two researchers are launching the Sustainable AI Group to promote environmentally friendly practices in AI development. They aim to provide organizations with tools and strategies to measure and improve the sustainability of their AI systems, amidst growing concerns about the environmental impact of AI technologies.
- Public discussions about AI sustainability have been overshadowed by the focus on computing power and geopolitical competition.
- Sasha Luccioni, formerly of Hugging Face, and Boris Gamazaychikov, ex-head of AI sustainability at Salesforce, are leading efforts to make AI sustainability practical and measurable.
- Their Sustainable AI Group will offer research-driven guidance and tools to help organizations understand and reduce the environmental impacts of their AI usage.
- Many companies still prioritize sustainability internally, despite the public discourse shifting towards competitive AI development.
- Luccioni highlights that AI systems can be designed with sustainability in mind, such as choosing smaller models or utilizing renewable energy for data centers.
- There is confusion regarding the water and energy requirements of AI data centers, with traditional cooling systems often wasting significant resources.
- Many enterprise applications do not require large models, and smaller, specialized AI systems can be more efficient and sustainable.
- Luccioni suggests organizations should assess their specific needs for AI rather than defaulting to large models, which can lead to unnecessary energy consumption.
- Market demand for sustainable AI practices may drive change in the industry, as customers begin to prioritize renewable energy and sustainability in their procurement decisions.
- Despite challenges, Luccioni remains optimistic about the potential for improvement in AI sustainability practices.
sustainabilityenergy consumptionenvironmental impactai boomresearch