How Michigan universities are at the forefront of advancing AI technology

With artificial intelligence continuing to expand, along with growing uncertainty about how it will impact the future, academic institutions stand at the forefront of advancing AI technology. In Michigan, professors at Michigan State University, the University of Michigan, and Wayne State University, are not only driving AI advancements through research and educational initiatives, but are also shaping its responsible implementation across various industries. 

The University of Michigan (U-M) in Ann Arbor has an AI Lab, which serves as a hub for collaboration, bringing together faculty and students to tackle AI-related challenges in areas such as image processing, natural language processing, and machine learning. 

“In academia, we really have the freedom to explore the variety of directions that are driven by AI, so we can do research that will pay off down the line maybe in a 10-year, 20-year horizon as opposed to the very short term,” says U-M Assistant Professor of Computer Science and Engineering, Michał Dereziński. 

Dereziński’s personal research focuses on reducing the computational costs associated with training AI models to make advancements accessible to more people while supporting “the next generation of data science algorithms.” 

“Only those with enough resources can actually build models [to advance AI],” he says. “My research essentially is focused on designing algorithms for doing this data processing that is required under the hood of training the machinery models so that hopefully we can train a better ChatGPT with reduced resources.” 

At Detroit’s Wayne State University, Professor Hengguang Li, Chair of the Department of Mathematics, pioneers AI research aimed at solving complex mathematical equations.

“There is tremendous potential using artificial intelligence algorithms to solve some of the most difficult equations,” Li says. “We develop AI to solve those equations because there's no existing model.”

His research seeks to develop AI algorithms tailored to solve equations unattainable through traditional numerical methods, bridging the gap between theoretical AI advancements and real-world applications.

In 2022, Wayne State received a donation for a campus-wide initiative to promote programs in data science and artificial intelligence. For this, the Math Department collaborates with the schools of engineering, business, and medicine to build up partnerships with students and faculty across campus. 

When it comes to academia, Li feels that AI will play a very important role in higher education for grading, assessment of a student's performance, and more. While he utilizes the benefits of AI, he also recognizes the risks, including academic misconduct of students with the availability of chatbot platforms.

“There are a lot of discussions on campus, but for now, people are still exploring different types of solutions to that,” Li says. “It is very challenging and there are a lot of risks to higher education in that way.”

He feels that while AI can be helpful, its widespread availability should urge people to think more about the value of education.  

“The perspective of people in terms of forward education will be different,” Li adds. “I think that's the challenge and also the opportunity for university leaders: how to embrace the change, especially from AI, and how to implement policies and adapt to the right technology to really improve what we do in higher education.”

In East Lansing, Professor Arun Ross at Michigan State University uses AI to develop a curriculum to help college students succeed, especially since students of unique backgrounds have different learning styles and needs. 

“We are trying to use AI tools to develop scenarios where we make some assumptions about the geographical area where the [learning] is happening or the background of the student,” Ross says. “We are able to use AI tools [to create] customizable tutorials, which the students can then utilize to better understand certain concepts.”

However, Ross also acknowledges that AI can be harmful if someone relies on it completely and doesn’t take into account context or other factors,  such as incorrect responses and biases. 

“AI tools are trained using data, sometimes historical data, and historical data can exhibit some biases, which have to be eliminated so that the resulting AI tool is not exhibiting those biases,” Ross says.

So, the work being done does not only include developing powerful AI tools, but also creating techniques to validate the data produced by AI. 

“We’re trying to see how synthetic data generation can benefit our students both in terms of teaching concepts but also in terms of them being able to utilize it for some of their deliverables,” Ross says.  

The educational role of universities is fostering AI literacy among students and end-users, he adds. By developing curricula and raising awareness about the potential risks and benefits of AI, universities have a crucial part to play in shaping responsible AI usage.

“Universities play a significant role in being able to educate individuals, companies, and other enterprises, about both the benefits as well as the harmful effects of AI, so that balanced decisions can be made,” Ross says. “As much as we develop AI technology, as much as we utilize AI technology, it is also important that we educate society about both the pros and cons of AI.”