What three Californians learned in Kansas about artificial intelligence | Opinion

As college instructors, we live with the sobering reality that some of our students use artificial intelligence for coursework, and in our home state of California, some colleges (including two of ours) are even providing their students with products such as ChatGPT. To become more AI-aware and less awash with worry, we recently attended AIDL, the AI and Digital Literacy Institute at the University of Kansas. The brainchild of Kathryn Conrad and Sean Kamperman of KU’s English Department and their partners at the National Humanities Center and the Hall Center for the Humanities, AIDL brings together educators from high schools and colleges around Kansas and other areas to confront the pedagogical, ethical and institutional implications of AI.

There, we grappled with understanding the opportunities and challenges of AI products in education through seminars, workshops and dialogue with leading scholars — an instructional model that provides a sharp reminder about the importance of trust and human relationships in the learning process.

Along with meeting experts in computer science, comparative literature, informatics, rhetoric and writing, AIDL attendees, in true humanities form, asked questions for which there are no easy answers. For example, we don’t know the long-term effects of offloading thinking and emotions to AI on students’ reasoning abilities or mental health. We don’t yet understand the implications of ceding critical elements of education to a small number of corporations that already play outsize roles in our lives.

We do know, however, the following about AI:

1. AI machines do not “think” or “understand” or “know” anything. They cannot empathize, as they have no feelings, and they are not objective “fact machines,” as they produce outputs prone to error and bias.
2. As sophisticated prediction machines, AI tools such as ChatGPT can produce plausible sounding text — that may or may not be true. Significantly smaller, task-specific AI models are being studied by researchers for applications such as drug discovery and weather prediction.
3. Large AI models are trained on billions of information fragments, running in data centers that demand tremendous amounts of fossil-fuel sourced energy and drinking water. A ChatGPT query might use more than 10 times more energy than a Google search, and one research team estimated that a series of five to 50 ChatGPT prompts may use approximately 16 ounces of water.
4. Artificial intelligence cannot run without humans. Writers and artists have had their work used to train AI without consent or compensation, and millions of workers are engaged in data annotation — the tedious task of labeling and sorting information to train tools like chatbots, while earning as little as $1 to $3 per hour.

Understanding what we know and don’t know about AI, educators at AIDL deftly ignored the hype and fearmongering. Instead, they scrutinized AI’s efficacy and ethics while putting students first. They designed a stunning variety of learning experiences: assignments that teach how AI works, lessons prudently integrating AI, and distraction- and AI-free work. As a result, we left Kansas feeling less adrift, confident that students at community colleges in Illinois, Iowa and Massachusetts, high schools in Georgia, Missouri and North Carolina, and other institutions across North America, are in skilled hands — and that we, too, could and would do better.

This remarkable experience was made possible by local donors who stepped in after federal funding was pulled. Kansans, it seems, understand that nothing replaces learning with others, sharing backyard cherries and rainy-day car rides. We are grateful and hope more educators can connect — to affirm our joyful human messiness, while tackling the essential questions of who we are and who we want to be with or without artificial intelligence.

This story was originally published July 3, 2025 at 7:04 AM.