Sasha Luccioni Discusses AI Costs at RISE Maui Summit

Serdar HocamAuthor & Editor

Sasha Luccioni, co-founder of the Sustainable AI Group, evaluated the environmental, social, and psychological costs of artificial intelligence at the RISE Maui Summit.

◉ 0 views
RISE Maui speaker examines costs of AI

Sasha Luccioni, co-founder and chief scientist of the Sustainable AI Group, shared the impacts of artificial intelligence on the environment, society, and human psychology with attendees at the RISE Maui Summit held at the Hyatt Regency Maui Resort and Spa.

Environmental and Carbon Costs

Sasha Luccioni contributed to the development of an open-source software tool called CodeCarbon to estimate carbon emissions from machine learning models.

She stated that a single data center can emit up to 4 million tons of CO2 per year, which is equivalent to the emissions of a small country like Iceland.

Water Consumption and Data Centers

A study revealed that for every 20 to 50 queries asked of large language models, the equivalent of a 500-milliliter bottle of water is consumed.

Morgan Stanley projects that global artificial intelligence data centers could consume more than 264 billion gallons of water annually by 2028.

Human Thought and Psychology

An MIT study showed that artificial intelligence models affect people's thinking patterns and result in lower levels of brain activation.

Attention was drawn to the phenomenon of AI psychosis, which excessive chatbot use can trigger and which could create a dangerous feedback loop.

Creative Arts and Copyright

While AI musicians, bands, and films create concern in the art world, Luccioni stated that artificial intelligence does not produce art.

It was emphasized that existing systems remix human creativity gathered from the web without permission or payment.

Hope and Solutions for the Future

Stating that companies are profit-driven, Luccioni conveyed that their good intentions alone cannot be relied upon and that true transparency is mandatory.

It was stated that a more sustainable technology can be built by ensuring transparency across the value chain and favoring smaller models.