Perspectives in Artificial Intelligence: Creating jobs, not replacing them | Marquette Today

A longtime writer and journalist might be the last person one would expect to embrace artificial intelligence, but that accurately describes Scott Rex’s outlook on the matter. 

“AI is a content production process, but it’s not thinking yet.” Rex says. “We want to teach students to use the AI engine to enhance their process by writing prompts that make the machine an aid to what the human is already doing. 
 
Rex, an instructor of practice in marketing and the co-director of the Northwestern Mutual Data Science Institute, has a hard job. He has to incorporate AI into his classes in a way that grows students’ proficiency but won’t become obsolete or circumvent the learning process. Ultimately, his goal is to get students jobs in the short-term while cultivating a set of skills that can’t be outsourced to AI for the long-term. To do that, Rex has incorporated data science and analytics into his toolkit.

Augment, Replace, Create 

When Rex teaches artificial intelligence, he talks about the ARC process, which stands for Augment, Replace and Create. In phase one, a new technology builds upon the existing capacity of workers, allowing them to do their jobs easier and faster. In the second phase, it replaces those workers. In the final phase, it creates new job for upskilled workers. 

Companies are looking for students skilled in artificial intelligence tools right now. However, Rex is cognizant that many businesses are looking to move to the “replace” phase. He is principally occupied with one question when designing his classes: how can students weather “replace” and move to “create?” 

He sees the solution in Marquette’s liberal arts core, which teaches students how to think, not what to think. 

“Learning how to reason through a process with creativity and critical thought is paramount,” Rex says. “The mental work you have to go through is the valuable part and you’ll be shortchanging yourself if you focus on the product over the process.” 

Training on the right data 

Most of the public hype around artificial intelligence has come from large language models trained on data scraped from the internet. This will inevitably lead to original works being used for unauthorized purposes and possible legal claims of copyright infringement. 

“As a former journalist and writer, that’s painful that my written work is going to be ingested to train an LLM that might be used by someone to try to replace me,” Rex says. 

Instead, Rex sees greater value in training artificial intelligence on internal, proprietary data. Companies could feed data from their customer relationship management database into AI software and identify trends much faster and with greater accuracy than before. 

“Retrieval-augmented generation AI is going to be huge in the next five years,” Rex says. “I may not particularly care what the whole internet has to say about something, but I want software that can query my own data to find connections.” 

A Challenge Engine 

Another role that AI might fill is that of the “challenge engine.” Rex encourages his students to ask large language models to give them ideas on how to respond to a prompt, then do independent research on the ones that sound most promising. 

This ensures the learning and critical thinking processes are still happening; they’re just happening more quickly than they would have without AI. 

“Artificial intelligence should open up different avenues of inquiry. In the past, you would have encountered things outside your frame of reference by serendipity; maybe in the library or through a random Google search. The value of AI is engineering those serendipitous moments and getting you on the right path more quickly,” Rex says.