Downside of AI - Lamar University Press

Examining environmental concerns surrounding AI usage

Almost everyone has heard of ChatGPT, Microsoft Copilot and Google AI. A good portion of those who’ve heard of these systems have likely logged onto these platforms and input a prompt or two. In fact, a simple Google search will bring up an artificial intelligence overview that provides a summarized answer to the search with links to validate the information.

AI is becoming the norm when using the internet, as it’s showing up everywhere. Generative AI can do many tasks aimed at helping to make the lives of human users more efficient and easier. Generative AI also aids students by helping them study by breaking down complicated terms into simplified terminology.

AI was made to make lives easier.

However, concerns revolving around the use of AI software have become more apparent. It’s important to understand what generative AI is and how it works.

What is AI?

Copilot, ChatGPT and Google Gemini are the most commonly used AI systems. However, what classifies as AI goes beyond those systems.

The National Library of Medicine defines generative artificial intelligence as, “A system of algorithms or computer processes that can create novel output in text, images, or other media based on user prompts.”

For example, imagine a person gives ChatGPT the prompt, “Create a pink computer wallpaper with hearts.” ChatGPT would not solely search the web for a pink wallpaper that met the requirements, but it would use the text in the prompt to generate a wallpaper itself.

Kimberly Conner, Lamar University director for university innovation, said the AI tool is not producing a new answer to a question.

“It's combing through the extensive data sets it has access to and then identifying patterns that would most likely respond to your question, whether it's just an answer or generating content or video or audio — but that's how it works,” she said.

An article published by Microsoft explained that generative AI systems are “built upon the foundation of neural networks.” These neural networks take inspiration from the human brain, and they are interconnected to exchange information.

Moreover, Microsoft notes that, “Generative AI uses advanced machine learning techniques to analyze large datasets and generate new content based on the context, style, structure, and tone of the original data.” Generative AI continues to grow in capabilities, with Microsoft noting that ChatGPT showed a breakthrough in natural language processing, such as showing capabilities in producing relevant and natural dialogue.

True or false

AI systems have started to become sophisticated, perfecting their way of working. They are blurring the lines between reality and fiction.

It isn’t uncommon to see at least one AI-generated video on your “For You Page” on TikTok. These systems can create videos, advertisements and even entire songs. More people are finding that the songs they’re hearing on social media are created by artificial intelligence software. As this technology grows, it’s becoming increasingly more difficult to determine if something is real or generated. Where AI-generated movements used to be obviously fake, the more AI is studied and updated, the harder the flaws are to spot.

There are multiple reasons why it has become increasingly difficult to tell the difference between what is AI and what is real. AI is a mirror image of human information, creating a digitalized format of information, Conner said.

“It's the digitalization of very old works of literature or art pieces or pieces of music, production, human content that could not have been digitalized at the time,” she said. “But subsequently, we have found ways to capture those in digital form. And so, the mirror — what it's able to reflect has become much more detailed and much more clear.”

Users must be wary about misinformation when it comes to social media posts. Fact checking and ensuring that what is being posted is accurate will help clarify if something is AI or not.

“It's difficult to tell the difference between human-generated and AI-generated, because we haven't had, in the past, to really sharpen our skills of discernment,” Conner said. “And that means developing the ability to cross-check information, to read laterally, to ensure that the information being presented to us is actually accurate.”

Positives of AI

AI is changing education, as it offers new ways of accessing information, Conner said, adding that it allows for tutoring or supplemental learning, and increases the ability to enhance digital accessibility.

“We are absolutely seeing and will continue to see higher education changing,” she said.

For example, if a student is struggling with a topic in class and needs help to break it down into smaller bits of information, AI is there to help, Conner said.

“AI has the capacity to aggregate information and find patterns for us that might help us solve problems or find resolutions to really sophisticated, complex problems that we haven't been able to solve in the past,” she said.

The use of AI is also changing how to plan more efficiently. AI can help support better medical diagnosis, forecasting models, infrastructure monitoring, and even improve the operation of a power grid, Thinesh Selvaratnam, LU associate professor of civil and environmental engineering, said.

“I'll give you an example of Southeast Texas,” he said. “Every time we get a hurricane, they have what they call a cone of uncertainty, because that's based on multiple models. Ideally, you want that as small as possible, so we can plan, right? AI can help. That's the benefit of AI. It's really good in crunching data, and it can come up with solutions.”

AI also aids in the environmental sphere, where it is used to help in preventive measures, creating a proactive approach.

“It can predict, ‘OK, there's going to be a leak,’” Selvaratnam said. “We already have models, but AI can accelerate and help us more. So, usually, when it comes to environmental engineering or the environment, it's not AI versus environment.”

What is important is that AI is used for good, without violating ethical standards such as class guidelines or copyrights, Selvaratnam said, adding that AI can also be part of a solution to help communities.

“It has the potential to help communities to make better decisions, respond more quickly to risk and operate infrastructure more efficiently,” he said. “It's really important for Southeast Texas. We talk a lot about resilience these days. AI can really help on that aspect.”

Negatives of AI

A criticism of AI-generated creative works has been the erasure of human creativity. An AI result produces an aggregation of everything seen on the internet that matches the search. Therefore, a person’s intellectual property can be used without recognizing it as intellectual theft.

“While the amount of data AI tools are able to access can be very helpful, we also have to be really mindful of using transparent systems that allow us to see where that data has been pulled from,” Conner said. “And we must continue to advocate for fair consumption. One way to think about that is ensuring AI tools we're using are not violating copyright and are not violating artistic creativity.”

An AI creation cannot replace human creation, Conner said. The value of human creation is held to a higher standard than an AI artwork.

“Personally, I would much rather have an imperfect human painting than an aesthetically ‘perfect’ AI-rendered painting,” Conner said. “So, (AI) capacities are being able to access information and iterate, to find, to get responses, or generate helpful artifacts. But the limitation is that we don't want to impinge on art or human creativity.”

Environmental impacts

While many words have been written about the intellectual impact of generative AI, less attention has been paid to the environmental impacts of the data centers needed to operate the systems.

Data centers are buildings that store the computer hardware necessary to complete AI’s functions, specifically that store and process digital information. They are not AI-specific, but with the growth of AI, they have been popping up more frequently and being built even larger.

According to Data Center Map, Texas is second in the country with 411 data centers. Bloom Energy reported that Texas is set to become the leader in data centers within the next three years. In the same period, Texas could see a 142% increase in its share of the data center market. There are several reasons for this increase, Selvaratnam said.

“(The) economic climate in Texas is supportive to industry in terms of taxation and in terms of looser regulations,” Selvaratnam said. “So, that's why we are seeing this massive boom in big tech companies. If you think about Austin, it's kind of like the next Silicon Valley. That's the main reason why we are getting more and more, because the regulations are not stringent compared to some other states, specifically, environmental regulations. And, most importantly, the state level government is pro-business — we don't have state income tax to start with, and the existing infrastructure is pro-industry.”

The environmental impacts from AI are not just from the software, Selvaratnam said. The impact is from the AI infrastructure.

“That's the thing we need to remember because AI depends on data centers, it needs to have service,” he said. “All those data centers need to have cooling systems. They need electricity, and they need water, and then they eventually need to replace their equipment.”

Water cooling systems aren’t the only issue with AI and its data centers. AI contributes to the electronic waste problem. According to a report by the UN Environment Programme, electronic waste materials from AI contains lead and mercury which creates a problem for everyday life.

“They can move into the air,” Selvaratnam said. “They can move into soil, and they can eventually move into water. So, it's not just a disposal problem. They can directly affect people who have worked with it, and also nearby families and communities.”

Selvaratnam said once these hazardous materials enter the water, it becomes a widespread problem leading to other issues.

“It's not just an environmental problem,” he said. “As soon as they enter into any of the streams, they become an economic burden. If you think about these metals, it's gonna directly impact your health. Then it’s gonna come with missing work, missing school. That's the economic problem.”

Transparency and planning are some of the most useful policies, Selvaratnam said. They would be clear rules rather than broad or punitive restrictions.

“The best practice of what we can do is clear disclosure of electricity and water demand,” he said. “Better coordination with utilities and grid operators, stronger standards for end-of-life equipment management, and incentives for efficient cooling, reclaimed water where appropriate, and lower impact facility design. So, the key is, when we are providing incentives think environment in your mind, rather than, ‘OK, we're gonna just give you a blanket tax incentive.’ Give the incentives that tie to environmental things.”

Being educated about AI allows users to use it responsibly in each situation, including the environment, Conner said.

“I know AI is a polarizing topic, and no matter how we feel about AI, the more we learn about it, the better we can leverage AI in ways that feel authentic, transparent, responsible and ethical to us, because it is a ubiquitous technology,” she said “It's not a technology we can ignore, because there are many times we are using artificial intelligence or platforms that include artificial intelligence technologies that we're not even aware of. You don't have to be a fan of AI to learn about it.”

Texas is facing a technological change with AI and data centers coming to the state in increasing numbers. Residents must be taken into consideration during expansion to ensure their livelihood isn’t being infringed on.

“Texas does not have to choose between innovation and environmental stewardship,” Selvaratnam said. “The real challenge is making sure growth in AI is matched by smart planning for water, power and waste. So, that's my key. Texas doesn't need to choose between, 'OK, we're gonna go with innovation or environmental stewardship.’

“We can make smart planning, so our innovation and stewardship matches.”