Mind Matters Natural and Artificial Intelligence News and Analysis
Students making notes
Image licensed via Adobe Stock

Learning to Communicate

Why writing skills are so important, especially in today's artificial world

Educators have been shaken by fears that students will use ChatGTP and other large language models (LLMs) to answer questions and write essays. LLMs are indeed astonishing good at finding facts and generating coherent essays — although the alleged facts are sometimes false and the essays are sometimes tedious BS supported by fake references. I am more optimistic than most. I am hopeful that LLMs will be a catalyst for a widespread discussion of our educational goals.

What might students learn in schools that will be useful long after they graduate? There are many worthy goals, but critical thinking and communication skills should be high on any list. I’ve written elsewhere about how critical thinking abilities are important for students and cannot be reliably faked by LLMs.

Writing Skills: No Longer Relevant?

What about communication skills, both written and oral? Some say that learning to write is no longer relevant. Just use an LLM. I disagree for three reasons. First, the actual process of writing helps us understand concepts and develop critical thinking skills. When we attempt to communicate an argument (or other message), we learn from our attempts to explain it clearly. After the initial draft, we revise—tightening our arguments, making our point of view clearer and more persuasive, adding additional supporting evidence, and anticipating possible objections. We learn from all of this. We wouldn’t if we asked an LLM to generate an essay for us.

The second reason why writing skills are important is that LLMs might do a passable job of analyzing Moby Dick or explaining Hume’s paradox of horror, but this is not what most students will be doing in their jobs in the real world. Some communications will be simple and perfunctory, like announcing a meeting, a revised policy, or a retirement, and LLMs might suffice. Communications that count are more challenging and the costs of botching them are more severe: a legal argument, a report on the results of a statistical study, an analysis of a proposed new business or government policy, a prediction about the future. Essays written by LLMs based on a scouring of historical text databases are likely to flop.

For example, Manfred Kiel leads a research team at Claremont McKenna College that makes regular economic analyses and forecasts for the Inland Empire, which consists of Riverside, San Bernardino, and other urban areas dozens of miles east of Los Angeles. Given all the hoopla about ChatGTP, Keil naturally wondered if he could hand his tasks off to it. He and colleague Sasha Rothstein asked ChatGPT to “write a 500-word economic analysis of the current economic situation in the inland empire 2023.” He did not ask ChatGPT to do an economic forecast which would have been far more challenging.

ChatGPT responded with an essay that included a lot of economic terms, like manufacturing, employment, and housing, but the economic analysis was mostly vague, superficial, or embarrassingly incorrect. For example,

The region has an unemployment rate of 7.4%, which is higher than the national average of 4.2%. The pandemic has had a significant impact on the labor market in the Inland Empire, with many businesses struggling to find workers.

ChatGPT had evidently found an internet mention of 7.4% unemployment in the Inland Empire and also found a separate internet reference to an Inland Empire labor shortage. These two statements were no doubt made at different points in time but LLMs struggle mightily with time. Not knowing what words mean, ChatGPT had no way of recognizing that its second sentence contradicted its first sentence. A labor shortage is extremely unlikely if the unemployment rate is 7.4%. In fact, the Inland Empire unemployment rate was 3.6% at the time of the internet mention of a labor shortage.

Another example. ChatGPT reported that high housing costs in the Inland Empire made it difficult for residents to live near where they work:

Another challenge facing the Inland Empire is housing affordability. The region has seen significant population growth in recent years, which has put pressure on the housing market….This has led to a housing crisis, with many families struggling to afford rent or buy a home. The lack of affordable housing is also impacting the labor market, as workers are struggling to find housing close to their place of work.

The real story is that many people choose to live in the Inland Empire and commute 1-2 hours to work precisely because housing prices there are much lower than where they work; for example, Orange County or downtown LA. Again, ChatGPT found some relevant factoids and strung them together in sentences that are coherent but wrong. After graduation, it is perilous for students to use LLMs to write things that are consequential.

Writing Bonds Us

A third reason why writing is important is that it bonds us. When I tell you what I think, you learn more about me. When you respond, I learn more about you. We learn about our similarities and differences and, if done politely, become closer. All of that is lost if our written communication becomes my LLM chatting with your LLM.

Oral communication is also important for students after graduation and shouldn’t be outsourced to BS-generating LLMs connected to text-to-speech algorithms. Students will need to talk to people individually and may be asked to make formal oral presentations; for example, summarizing the kinds of written reports described earlier. To do that well requires practice and the place to learn is in school where the stakes are low. In addition, just as the process of writing helps develop critical thinking skills, so does the process of speaking. Good oral presentations require that we know what we are talking about.

What kind of educational environment nurtures critical thinking and communication skills? There are many possibilities. One is to have students work in small teams on out-of-class projects, hand in written reports, and make oral presentations. Another possibility is a major assignment that requires students to make oral and written reports of their progress and culminates in a long-written paper and an oral summary. None of this can be fudged by LLMs. All of this can help students develop skills they will need after graduation.

Gary N. Smith

Senior Fellow, Walter Bradley Center for Natural and Artificial Intelligence
Gary N. Smith is the Fletcher Jones Professor of Economics at Pomona College. His research on financial markets statistical reasoning, and artificial intelligence, often involves stock market anomalies, statistical fallacies, and the misuse of data have been widely cited. He is the author of dozens of research articles and 16 books, most recently, The Power of Modern Value Investing: Beyond Indexing, Algos, and Alpha, co-authored with Margaret Smith (Palgrave Macmillan, 2023).

Learning to Communicate