LSU faculty, researchers explain pros and cons of AI in the state

AI data centers and the energy required to power them are growing concerns for Louisiana residents as large companies like Meta and Amazon continue expanding in the state.

The Louisiana Public Service Commission approved a new initiative in December that could cause utility ratepayers in the state to pay for much of the cost of powering these data centers.

The Lightning Amendment speeds up the process for utility companies looking to build power plants for large, energy-intensive customers. The Lens, a non profit newsrooom in New Orleans, says that many of these customers are AI data centers.

Meta began building of its Hyperion AI data center in Richland Parish, which totaled 2,500 acres, in late 2024. Fortune reported last month that Meta purchased 1,400 more acres adjacent to the existing plot. These plots of land combined are more than twice the size of Louis Armstrong New Orleans International Airport.

Louisiana Economic Development announced in late last year that a $10 billion Hut 8 AI data center will go in West Feliciana Parish. The Shreveport Times reported last month that Amazon selected Caddo and Bossier Parishes to build $12 billion data centers.

Brian Snyder, a professor of environmental sciences at LSU, voiced his concerns with AI data centers in Louisiana.

“How much energy are they using? Where are they getting the energy? How is it going to affect consumers?” Snyder asked.

Some residents of Richland Parish do not use AI at all, yet they might have to pay some of the costs for the data centers, Snyder continued. He added that companies like Meta often put data centers in rural, low-income parts of the country.

“Some of those places might want the jobs that come from that, but the people should get a say, in my opinion, about whether they want the economic development that comes with this,” Snyder said.

While the growth of AI and data centers poses threats to communities and the environment, many researchers and scientists also use AI for good.

LSU professor of petroleum engineering Mayank Tyagi uses AI to address problems in the petroleum and clean energy industries.

Tyagi said he has been interested in data-driven modeling for the last few years. He said he is currently analyzing data for carbon capture and storage projects.

He is looking at data from the field and trying to reconstruct seismic image data using generative AI. Seismic image data helps researchers locate and monitor sites for carbon capture and storage.

Tyagi explained that he also incorporates AI into challenges with offshore drilling operations. He said that one risk of drilling is gas kicks. Tyagi developed a gas kick detection machine, which is learning techniques known as sequence predictors.

Tyagi also uses AI to create algorithms that help optimize drilling.

“Basically, when you do the drilling, not only do you want to drill safely, but you want to do it efficiently, so it can save some money for the company and things like that,” Tyagi said.

Supratik Mukhopadhyay is another LSU professor who uses AI in his work. Mukhopadhyay, a professor of environmental sciences, incorporates AI into drug discovery and wildfire prevention.

Tyagi and Mukhopadhyay both have joint appointments in the LSU Center for Computation & Technology. Both professors recognize the benefits of AI as well as the downsides.

Mukhopadhyay said he used AI to help develop treatments for COVID-19, and he used AI in his work with antibiotic resistance. Mukhopadhyay added that he has also began using AI to discover drugs to treat ALS.

“AI is a transformative technology. It is changing the way our civilization operates,” he said.

Mukhopadhyay mentioned that humans have been trying to find a cure for diseases like ALS and cancer for decades.

“Even today, if you have one of these diseases, it’s a death sentence,” Mukhopadhyay said. “So since we haven’t found anything with our current techniques, there is a chance we have to use something better. That better thing is probably AI.”

However, Mukhopadhyay said there are some risks that come with using AI. It might make mistakes or provide biased results if the data it is using is biased. AI also has negative environmental impacts because of the energy and water consumption of data centers.

Tyagi elaborated on power generation for AI data centers. He said data centers need a highly reliable power source with no downtime.

At first, there was a push for renewables like solar and wind to power data centers, but those are intermittent energy sources, Tyagi explained. Solar only works for half of the day, and wind only works when there is wind.

Tyagi said AI data centers mainly use natural gas because it is currently the most reliable electricity generation source in the United States.

He explained that some states have strict regulations on emissions or not enough natural gas resources. The oil and gas industry and state policies in Louisiana are favorable for these large data centers.

Tyagi thinks communities that house data centers, which might not benefit from the data center at all, should have a say in decisions.

“So should they be on the main utility grid versus their own micro grid? That’s a very good question,” Tyagi said.

Tyagi educates his students on how to properly use AI. He believes individuals can incorporate AI into their work and thrive.

However, he said these large data centers should make sure they are solving more problems than they are creating.

“My message here that remains is all these deep-pocket companies like the AI data centers have a social responsibility. They have huge environmental responsibilities,” Tyagi said.