The New Orleans City Council enacted a year-long ban on data centers on Jan. 28 after the cancellation of a proposed center in New Orleans East. Citing mistrust and a lack of transparency for the development, the city council voted unanimously to ban the center’s construction.
The New Orleans Comprehensive Zoning Ordinance does not have a clear definition of data centers, making it difficult to ensure proper development and safety, according to Zoning Ordinance Councilmember Jean-Paul “JP” Morrell.
AI data centers have existed for decades, often in rural areas, though they have evolved considerably since the introduction of the Electronic Numerical Integrator and Computer, which entered service in 1946 and was considered one of the first data centers. As the internet expanded in the 2000s, data center construction exploded, eventually leading to the birth of large-scale data facilities in 2010.
“Data centers are where the internet lives … We tend to think the internet and data as being kind of ethereal and in the air floating around,” Tulane University Nathaniel Rich, professor of practice of environmental studies, said. “But actually, the internet lives in computers called data servers … in these giant warehouses across the world.”
As AI centers move in, Tulane researchers have raised red flags about energy use and environmental impact.
Data centers require enormous amounts of cooling by fans and air conditioners because they give off so much heat, making them highly energy intensive. New Orleans is particularly vulnerable to environmental hazards due to its coastal location.
“There’s much more awareness of environmental risk and a wariness to add big industry here,” Rich said. “You would also think the company wouldn’t be eager to embark on a huge industrial project in a place that’s so vulnerable.”
Rich said in the race to become the fastest-growing tech companies, firms are rapidly building AI technology centers with little regard for community input or environmental concerns.
“[Developers are] trying to put it everywhere … the companies are all competing against each other,” Rich said. “A lot of land in parts of the country that are very, very poor and out of the way are suddenly having this gold rush.”
Expanding wherever it’s cheapest, Rich said data companies often overlook broader environmental and social consequences when deciding to expand into more populated cities.
“Some of these [data centers] are starting to come a little closer. You’re starting to see more pushback in a lot of communities,” Nicholas Mattei, associate professor at the School of Science and Engineering, said. “It’s kind of weird to put them in Louisiana, because we’re not really on a main trunk for a lot of the data infrastructure.”
As AI becomes a more significant topic of conversation, Mattei said that Tulane University can play an important role in shaping how these technologies engage with local communities.
As home to the Center for Community-Engaged Artificial Intelligence, “we’ve got unique partners in the Gulf Coast area and in New Orleans,” Mattei said. “I think we can really lead in showing how to effectively partner with nonprofits, local communities … to make sure that those benefits from AI are done in conversation and co-created with those communities that are impacted.”
Tulane’s chapter of Epsilon Eta, a pre-professional environmental fraternity, engages with local environmental issues through service initiatives and volunteer opportunities.
“This [cancellation] is a win for communities, and specifically environmental in the environmental justice space,” junior Epsilon Eta Programming Chair Georgia Hoffman said. “New Orleans East got a really big win for this [the cancellation], especially since that is an area of the city that… often has to deal with the most environmental hazards and harms by development.”
Despite the drawbacks, Mattei said data centers are becoming increasingly necessary as AI use grows.
“Do you plan on using technology less, going forward? Am I going to stop talking to my phone? Am I going to stop streaming music and stuff like that? Probably not,” Mattei said. “So, there’s a demand there, and they’re trying to meet it, and it needs to be met by some computational power somewhere.”