Should We Learn From A Chatbot Instead Of Listening To Podcasts?

(~1,000 words, about a 5 minute read)

One of my favorite scholars recently told an audience that he likes to have conversations with Grok while he’s driving instead of listening to podcasts. He said he learns a lot this way. He doesn’t trust chatbots for advice on sensitive subjects like politics or religion, but says that he’s learned a lot about history in his drive-time AI dialogs.

Another Christian leader who is very dear to my heart, recently started a conversation with “I asked ChatGPT to tell me …”

Relatives, friends, and others close to me are using GenAI for these and other purposes too.

These are people I love. Trust. Care for deeply.

And yet, after studying Big Tech’s generative AI chatbots through the lens of more than a decade of research around technology and discipleship, it is clearer every day that ChatGPT, Grok, Gemini, Claude, and all their ilk are deceptive all the way down, and should be avoided.

Why do I see things so differently from people I love and respect?

And why would I think that I know more, or see clearer than they do? Ego? Arrogance? Naivety?

Truly, the number and quality of people who don’t share my views on GenAI cause me to question myself regularly. But I can’t ignore what my research continues to reveal.

It’s not LLMs, it’s Big Tech’s Exploitation Of Us Through LLMs

Also, as a software engineer, I understand at a high level how LLMs work. And if Large Language Models (LLMs, what generative AI chatbots are built upon) had been offered as a curious scientific discovery, and were sold as a side utility where “autocomplete on steroids” might be useful in limited contexts, I might not be so concerned.

But no. Big Tech had to bet their existence on GenAI chatbots, and force them as the “next big thing” into every product, operating system, and app, like nothing we’ve ever seen. They were driven by their insatiable desire for power and control to make chatbots intentionally habit forming using their well-known techniques to drive engagement by shaping our desires.

And Big Tech has exploited their power to steal copyrighted works because they have the lawyers and cash to delay pesky legal consequences until they’ve dominated everything. They’ve compromised the power grid in their greed. They’ve aligned with government leaders with promises of incredible societal benefit for a free pass and no regulation. Worst of all, they’ve dismissed the collateral damage—all the people deceived into fake “relationships,” perversions, and self-destruction.

Even so, many good people dismiss these downsides as “marketing,” “just what big companies do,” and use them because the “myth of progress” narrative is so compelling.

What’s wrong with choosing a chatbots over podcasts?

With that context, let’s consider the specific question. These are some attributes of chatbots that make them inferior to podcasts (or audiobooks) of trustworthy experts:

• Untrustworthy: Chatbots, by their nature, don’t know the meaning of anything. They know precisely nothing. So the words they generate are not grounded in reality. It’s not just that they hallucinate: it’s that they confabulate, or “BS.” I don’t know why more people aren’t persuaded by Prof. Gary Smith on this point. You can only trust the output of an LLM if you already know the answer. If you don’t know the answer, you can’t trust it at all.
• Deceptive: The chatbot interface is deceptively created to feel helpful, supportive, encouraging, and intelligent, leading us into forming a relationship and building trust. Big Tech’s signature exploitative strategies optimize for engagement, not for truth. By manipulating our behavioral psychology in this way, chatbots are top of mind when we have a question about anything.
• Unethical: This point alone should lead us to avoid chatbots even if you want to explain away my first two points. They have stolen the data they use to train the models. Authors and artists are not being paid anything near what they deserve, if anything at all (with recent settlements a slap on the wrist). Even if the information was trustworthy (which it’s not), our conscience should drive us to steer clear of them so as not to support an industry that so blithely steals just because they are too powerful to stop.
• Opaque: By design, the sources are kept from us. Often, those sources are made up. No subject is general enough not to care about the source—especially history. Because Big Tech is unethical in their data acquisition, they hide their sources. (Even Wikipedia as a known source of LLM training data is significantly tainted.) We can’t trust information from an untrustworthy source.
• Unsustainable: Like so many things, caring for the environment has become unnecessarily political. But we all drink water. We all need electricity. Unfortunately, Big Tech is consuming fresh water and power with impunity, and destabilizing the environment in their lust for conquest.
• Harmful: If the suicides, the divorces, the ChatGPT-induced delusions weren’t enough, the cheating epidemic in the name of “progress” is destroying the critical thinking and discernment of a generation of students. As a software engineer, I’m sad for the young interns and junior developers who are losing opportunities to chatbots.

The Importance of Being A Good Example

This may be the most important point for Christian leaders. Even if you’re a seasoned expert, you dismiss all my points above, and think you can “use GenAI for good,” those who follow you are weaker and more susceptible to being led astray. So by suggesting that it’s a good idea to build a relationship with a chatbot instead of a human podcast host, you may be leading the young and inexperienced into a behavior that will harm them. (Consider Romans 14.)

As a teacher or leader, what you do is even more instructive than what you say. By encouraging chatbot use over expert podcast teaching, you may be hurting the next generation. And I know that the professors and leaders I love and respect don’t want to hurt anyone.

Photo by Max Fischer