Concerns about the environmental impacts of advanced technologies such as AI is prompting debate over whether computing-intensive applications and the chips that power them need to be regulated.
That’s according to experts speaking during the “Advancing Technology for a Sustainable Planet” conference hosted by the Stanford Institute for Human-Centered Artificial Intelligence and the Stanford Woods Institute for the Environment.
Technologies including artificial intelligence and cloud computing use energy that result in carbon emissions. At the same time, many of these technologies also help companies meet sustainability goals — meaning companies need balance between quickly adopting and scaling emerging technologies and understanding how those technologies could affect the company’s overall environmental impact.
Environmental impact depends on scale — particularly for a technology like artificial intelligence, said Peter Henderson, a Ph.D student in computer science at Stanford University, during a conference panel session. Companies often optimize AI algorithms to address energy use and carbon emission concerns before deploying a machine learning model.
Peter Henderson Ph.D student, Stanford University
“The point is to make sure we don’t scale to a point that is harmful to the environment, when the goal of a lot of machine learning work is AI for social good, where we want to build more sustainable things, we want to optimize batteries, energy grids,” he said. “But if all that optimization leads to more negative impact than positive, it’s not worth undertaking.”
Henderson said beyond company measures to optimize such technologies, government efforts to provide rules for AI usage will likely also be necessary. While some efforts to regulate AI are already underway in the European Union, they don’t fully address the environmental impacts of the technology.
Targeting AI’s environmental impact
The EU’s AI regulations aim for consumer rather than environmental protection, according to Henderson.
He added that a significant piece of AI’s environmental impact comes from GPUs powering the AI and machine learning models. Regulation would have to cover chips and other technologies underlying AI use to address environmental impact concerns, he said.
“Here in California, there was recent regulation [that says] some GPUs are not allowed to be sold here anymore because they’re not efficient enough,” Henderson said. “That could be a path forward to pushing innovation and forcing more efficient chipsets.”
Though incentives — such as lower cost — exist for companies to use more efficient chips, Henderson said it is still an area that needs to be considered from a regulation perspective.
A lack of data on the environmental impact of technologies such as AI, cloud computing and bitcoin makes effective regulation difficult, he added.
“Step one is making sure we have enough reporting, enough data to make good regulatory and policy decisions,” he said.
Salesforce exec talks sustainability
Measuring a company’s total carbon emissions is a difficult task, with a significant amount of reporting traditionally being estimates rather than actual measurements, said Kathy Baxter, principal architect of ethical AI practice at Salesforce, who also spoke during the conference.
“There’s no way for us to really know are we truly getting better, and being able to see, what is correlation, what is causation, being able to identify,” she said. “If you don’t know where your emissions are coming from, you can’t possibly control them.”
Salesforce has been actively pursuing sustainability goals, launching its Net Zero Cloud as a way for the company and its largest customers to measure Scope 1, 2 and 3 carbon emissions for the entire supply and value chains. Scope 1 and 2 emissions are directly controlled by a company, while Scope 3 emissions occur outside a company’s control, such as from companies within its supply chain.
Baxter said more data is needed to help further sustainability efforts.
“There’s no way any single company, any single government, is going to be able to solve this problem,” she said. “We’ve got to pool our data together and we’ve got to figure out the solutions together.”
Makenzie Holland is a news writer covering big tech and federal regulation. Prior to joining TechTarget, she was a general reporter for the Wilmington StarNews and a crime and education reporter at the Wabash Plain Dealer.