By: Margarita Garcia - Managing Director, Naoitech
Times are changing and we must evolve with the realities of today. AI is advancing at a steady pace, but are we missing again the targets and purpose of true technological innovation? Technological innovation is not about recreating human intelligence or mimicking what humans can do well, it is about advancing areas that can enhance society as a whole.
Responsible innovation is a framework that ensures that technological advancements are developed and deployed in ways that benefit society as a whole. It involves a proactive approach to identifying and mitigating potential risks and negative consequences. The current landscape of the AI industry is concerning as some of the AI companies and projects are not taking into consideration the impact vs value equation. While LLMs have undoubtedly pushed the boundaries of AI, their widespread application in various domains has led to a certain degree of homogeneity. Many AI products and services are now built on similar foundational models, offering incremental improvements rather than valuable innovations. We have entered a race of market valuation rather than societal significance, however, this time the careless development of AI technology and the gamification of AI can create irreversible problems for our environment and future generations.
According to a study by Researchers at the University of California, Riverside “Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models” AI is wasting massive amounts of water and electricity. It is estimated that using GPT-4 to generate 100 words consumes up to 3 bottles of water and generating an image on AI uses the power to charge a smartphone. In their paper, the researchers highlight the importance of developing sustainable practices that address the water and carbon footprint of AI.
The growing carbon footprint of artificial intelligence (AI) models, especially large ones such as GPT-3, has been undergoing public scrutiny. Unfortunately, however, the equally important and enormous water (withdrawal and consumption) footprint of AI models has remained under the radar. For example, training GPT-3 in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand may be accountable for 4.2 -- 6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4 -- 6 Denmark or half of the United Kingdom. This is very concerning, as freshwater scarcity has become one of the most pressing challenges shared by all of us in the wake of the rapidly growing population, depleting water resources, and aging water infrastructures. To respond to the global water challenges, AI models can, and also must, take social responsibility and lead by example by addressing their own water footprint.
AI should be used carefully and with a conscientious mindset. It is important to start educating organizations and consumers about the responsible use of AI tools so collectible we can understand that using AI as a casual conversation partner, performing unnecessary, frivolous or repetitive queries or developing AI projects without real value can cause significant problems in our environment.
At Naoitech, we have conducted extensive research on numerous companies and strategies to evaluate the business value of initiatives. Our goal is to identify the best technology, providers, and processes for implementing projects that can boost revenue, reduce costs, and enhance operational productivity, all while prioritizing energy-efficient models and hardware. This may involve recommending process improvements, SMLs, AI Evaluations, or even traditional digital platforms.
The race to develop numerous LLMs, AI assistants, and Image Generators that perform the same tasks, along with the gamification of AI can have serious implications for the environment. The AI industry must analyze carefully the societal value of its products and services. As AI technology continues to advance, it's crucial to develop more valuable and sustainable solutions that minimize the environmental impact of these models. If as an industry we don’t take this matter seriously, we may be responsible for depleting our planet of the most vital resource for the existence of life.
Sources
Ren Shoalei, Li Pengfei and Yang Jianyi, and Islam Mohammad A. “ Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models” April 6, 2023
https://arxiv.org/abs/2304.03271
Danelski David “AI programs consume large volumes of scarce water.” April 28, 2023
https://news.ucr.edu/articles/2023/04/28/ai-programs-consume-large-volumes-scarce-water
Alexandra Sasha Luccioni and Yacine Jernite, Hugging Face, Canada/USA
Emma Strubell, Carnegie Mellon University, Allen Institute for AI, USA “Power Hungry Processing: Watts Driving the Cost of AI Deployment?” October 15, 2024