AI Divergence

Saturday-Sunday, September 21-22, 2024

AI Divergence

Alice Evans is a researcher, podcaster, and author studying how cultures differ in defining/assigning traditional male-female gender roles. Her Substack’s About page is at the link.

Noodling on the theme of AI brought me to her post on how universities should integrate AI into teaching. Starting into the post, which is more of a loose list of considerations, was this interesting tidbit:

“Before we dive into solutions, let’s consider the consequences of inaction. If universities merely ask students to promise no plagiarism, and otherwise continue with business as usual, we create four significant risks:

1. Falling behind: Some students may disregard AI altogether, leaving them ill-equipped for a labour market that increasingly rewards technological mastery. A new survey of 100,00 workers in Denmark finds that, even within occupations, women are less likely to use ChatGPT. This puts them at a disadvantage.”

She includes an illustrative graph from the Danish study at the link.

The differences between men and women in AI use are significant, but not overwhelming. Thus, you might wonder what exactly the statistics might tell us. For Evans, they tell us that university education (where she works) should integrate LLMs into the curriculum as a part of occupational training as well as for the purpose of figuring out how to prevent LLMs from becoming tools for students to complete homework assignments by cheating.

Statistics tell us a lot of things at once, which is to say they can be quite ambivalent. The gender divergence in Denmark might tell us things unique to Denmark’s culture, or they might tell us something about men being more willing to experiment with a new technology than women are. The latter would be in keeping with generally observed gender differences: men are prone to riskier behaviors than women, hence one might expect this to be the case even when it comes to new technology that doesn’t seem distinctly hazardous to physical health and safety.

Another question raised by the chart was what divergence might turn up if you studied people’s rates of AI adoption by age group. My suspicion is that the young would more quickly adopt AI and would be more open to playing and experimenting with it as a way of learning its capabilities. Children are more apt to learn a new technology intuitively by playing with it without inhibitions. And for an age group in higher education, the competitive crucible of academia would be a place where a lot of unguided experimenting goes on, I would suspect, as a means of perfecting cheating skills.

Nonetheless, Evans’s point that academia needs to figure out how to serve their students by recognizing AI’s significance as a new tool would seem inescapable. If higher education is a form of job training, for all intents, it should train students in using the latest tools effectively.