The progressively powerful artificial intelligence (AI) models are becoming increasingly hungry for electrical power. The carbon footprint of generative AI, which includes models like Gemini and ChatGPT, is overwhelming. We should act now before it snowballs into another challenge to the planet’s sustainability.
There is a need to balance technological advancement with ecological well-being. We can still exist if generative AI does not perform its wonders—but not if Mother Nature can’t survive the human-created energy and water demands of unsustainable technology.
To no surprise, sustainability is still not being considered in modern innovations. Social media is adversely impacting young adults and proving to be socially unsustainable. A small survey done in a college requires approvals from institutional review boards (IRB). Ironically, a huge social experiment in the form of social media required no approvals.
Take traffic—it continues to be a major impediment to protecting our environment. Road infrastructure and traffic engineering were unsustainable inventions that are obviously unscalable beyond a point. We cannot keep building roads infinitely to meet the needs of a growing population. Nor can the traffic be re-engineered easily to meet the needs of growing commuters through the decades. We cannot also expend energy indefinitely for our daily commutes.
The inventors, the public, and governments did not have this vision of the future when they unleashed this slowly progressing climate disaster. It is understandable if people did not foresee what vehicles would create when they were first invented. But when the repercussions were evident, governments and interests made overt attempts to oppose public transport and favor cars or remained indifferent. We cannot let artificial intelligence take the same route.
There are some ethical concerns with AI as well. In my recent talk at Stanford University, I discussed some of the threats AI poses to social and environmental sustainability. As I indicated in the video from the event, the unexplainable nature of the behavior of the current AI architectures is certainly a cause for concern.
It is imperative that the regulation around artificial intelligence address not only the social impact but also its environmental impact. Efforts have started but are currently not enough. There is an urgent need to fund projects to invent paradigms that can reduce the carbon footprint of AI models. In their current form, the underlying AI infrastructure will continue to need more power. Resourcing the energy needs from nuclear fusion as is being touted by OpenAI‘s Sam Altman is not enough. AI’s frameworks need to change.
It is not just energy that artificial intelligence is hankering for. The machines on which AI models run and the data centers they are housed in need tremendous amounts of water for cooling. Scientists estimated that by 2027, the water needs of AI models could reach that of half of the United Kingdom.
The needs are unsustainable. In addition, the e-waste and toxic material that the machines generate can cause considerable damage to the environment if not treated properly. Ecojustice must be factored into the design and deployment of AI systems.
We should invent tools and techniques that bolster sustainability rather than hamper it. Many things need to happen to ensure this. To start, government funding agencies such as the National Science Foundation must mandate ecological impact assessments to be submitted with every major grant proposal before funding projects. With time, the same must be extended to private funding as well.
Education must increasingly focus on sustainability. Some universities are already doing this. The machine learning course that I teach to graduate students won the Sustainability Across the Curriculum award from the provost of San José State University in spring 2023. It is high time that we pay attention to the very sustainability of life with every path-breaking invention.
Vishnu S. Pendyala, Ph.D., MBA (Finance), teaches machine learning and other data science courses at San José State University and is a Public Voices Fellow of The OpEd Project.
The views expressed in this article are the writer’s own.
Uncommon Knowledge
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
The post Will AI Be Another Unsustainable Environmental Disaster? appeared first on Newsweek.