How AI is transforming research and learning

Assistant professor discusses limitations, encourages use in post-secondary spaces

In the ever-changing landscape of technology, Artificial Intelligence (AI) has steadily gained prominence, particularly in the spheres of education and research — reshaping conventional practices and offering novel opportunities within these domains.

As society adapts to the evolving impact of AI, U of M assistant professor at Asper school of business Wenxi Pu shared insights into the intersection of AI, education and research. As an assistant professor specializing in management information systems, Pu discussed his role in teaching programming for business analytics, specifically in the MIS 3020 course offered by the U of M.

In this class, students learn Python programming and data analysis — incorporating open-source large language models, such as Llama 2, in coding and data analysis projects.

Last semester, Pu taught an introductory course focusing on business analytics where students delved into data visualization and the ethical use of large language models, emphasizing the responsible application of AI tools.

“Essentially, in my teaching, I encourage my students to use AI to help with their work,” he said.

He expressed the importance of responsible AI use, encouraging students to leverage AI for efficiency while remaining mindful of its potential downsides.

In his research, Pu focuses on three major perspectives of AI. The first aspect explores AI’s role in the hiring process, particularly in the potential biases it may introduce. Pu discussed the discrimination individuals with mild mental health challenges face, and how  AI-generated hiring processes might inadvertently disadvantage them.

“We are still trying to prove that humans actually cannot detect the differences because of mental health challenges,” he explained, highlighting his ongoing experiments designed to ascertain whether human evaluators could detect such differences.

The second perspective Pu’s research examines is human reactions to AI threats, examining how individuals can defend against job displacement by AI. He focuses on the concept of moral identity — the extent to which ethical considerations like compassion, justice and generosity play a role in an individual’s self-perception. Pu explores whether people with a higher moral identity are better equipped to navigate potential threats posed by AI.

“We believe that with this AI threat in place, individuals would have broader moral scope,” he hypothesized.

Pu also shared his use of AI in research related to climate change. Using text mining techniques, he analyzes millions of news articles to identify patterns and narratives about climate change across different countries. By leveraging AI and analytics, Pu seeks to uncover how climate change is discussed globally, considering sentiments and regional perspectives.

Pu shared his insights into the future of AI in both research and education, emphasizing the undeniable permanence of AI and its pivotal role in various aspects of people’s lives.

Focusing on the implications for education, Pu encouraged schools and universities to integrate AI into their programs and curriculums. He also stressed the need to effectively train students to leverage AI within their specific domains, and to be aware of the impact and shortcomings of AI.

“There’s a saying, ‘lawyers won’t be replaced by AI, it’s lawyers who use AI who will replace lawyers who don’t,’” he joked.

Shifting to research, Pu touched on the ways AI is shaping the future of scholarly pursuits. He highlighted AI’s potential to level the playing field for researchers with diverse language backgrounds. By automating language correction and streamlining processes, AI could facilitate more equitable participation in global research endeavors.

Pu outlined the different stages of the research process where AI can contribute, including idea generation, literature review, experimental design, data collection and writing. While acknowledging AI’s capability to assist in these areas, Pu also raised concerns about its limitations, particularly in terms of creativity. He cautioned against over-reliance on AI, emphasizing that it remains a tool to augment human capabilities rather than replace them.

While AI is expected to bring about significant advancements and efficiencies, Pu emphasized the irreplaceable role of human intellect, creativity and self-awareness in driving innovation and breakthroughs in these domains.

“I believe humans are still the drive of the creativity of new research breakthroughs,” he said. “I can’t tell what’s going to happen in the long term.”