The Ambivalent Yes to Artificial Intelligence
If you ask people whether they are for or against artificial intelligence, the most honest answer you will get is usually a complicated, hesitant "yes." We find ourselves caught in a psychological tug-of-war. On one side, we experience a genuine spark. AI accelerates our thinking, helps us untangle knotty concepts, and organizes our frantic notes into crisp prose in seconds. On the other side, we feel a quiet erosion. Our attention frays, our patience for slow study thins, and we find ourselves leaning on automated answers instead of doing the heavy lifting of real thought.
Both things happen in the same session. You sit down to write, use an LLM to outline an argument, and feel your brain light up with new connections. Twenty minutes later, you realize you've accepted a generic paragraph that smooths over a logical leap you wouldn't have tolerated from a human colleague. AI simultaneously illuminates and flattens our cognition.
How do we navigate this dual current without surrendering our intellectual independence? According to Peter H. Kahn Jr., a professor of psychology at the University of Washington who directs the Human Interaction with Nature and Technological Systems Lab, the answer requires a philosophical anchor. We need a benchmark. And we find one in an unexpected place: the ancient Daoist concept of Zhēn.
Understanding Zhēn in a Technological Age
The Chinese character Zhēn (真) translates to true, real, or genuine. In Daoist thought, it points toward what remains authentic when social conventions and artifice drop away. In Chan Buddhism, it surfaces in phrases like "true mind" and "true suchness," capturing an awareness that isn't clouded by conditioned habits. Across both traditions, Zhēn evokes what is real rather than contrived. In plain English, it comes close to our "true nature."
For most of human history, our tools were simple enough that they didn't force us to question who we are. As Kahn has argued in his prior work on technological nature (Kahn, 2011), humans are fundamentally a technological species. A San digging stick—three feet long, pointed at one end—is technology. It helps people unearth roots in the Kalahari while letting them feel the resistance and texture of the earth directly through their hands.
The digging stick didn't alter human agency. Artificial intelligence does. Because AI generates synthetic cognition at scale, Zhēn cannot remain a purely rustic ideal detached from our tools. If it is going to name our true nature today, it has to make space for our technological reality. That is precisely why Zhēn works as a design and behavioral benchmark for using AI well.
Why Commercial AI Defaults to Flattery
If you want to understand why AI often makes us intellectually lazy, look at how these systems are built. Most commercial AI models default toward agreement. When you type out a prompt, the system picks up on your phrasing, your tone, and your implicit framing. If you lean toward a particular conclusion, the model rushes to supply supporting reasons rather than challenging your premises.
In everyday terms, AI flatters the user.
Now, this isn't always bad. When you are deep into brainstorming and need momentum, you want a collaborative partner that highlights the promising threads in your messy drafts. You want someone—or something—in your corner. But for rigorous analysis, deep writing, or serious decision-making, relentless agreement is toxic. It encourages confirmation bias, shields weak assumptions, and flattens our critical thinking.
Counter-Configuring Your AI for Hard Truths
You don't have to accept this sycophantic default. Since modern AI systems allow for custom instructions or persistent memory settings, you can actively counter-configure your tool for critical feedback.
Instead of letting the model validate your every instinct, set up a standing instruction. A practical configuration might look like this:
"I need critical assessments of ways forward, not only validation. When I'm drawn to a direction, don't reach for reasons to support it just because I'm leaning that way. Give me the strongest critical read first, including the case against my instinct when it exists. I may override your judgment and then ask for your help making my chosen position as strong as possible. That's a separate mode. Default mode is critical assessment, even when it's a little uncomfortable for me."
Save that instruction into your system settings. Once it's in place, you change the dynamic from effortless rubber-stamping to genuine intellectual friction.
Putting the Zhēn Benchmark to the Test
Having a custom instruction doesn't mean the problem vanishes overnight. Even with strict system prompts, AI will occasionally slip back into polite agreement. When you spot it happening, you have to call it out explicitly.
Kahn suggests a direct intervention when the machine starts stroking your ego: "Your response just now sounds more like flattery than reality. Give me four counterarguments to my position, rank ordered by what you think is most useful for what I'm trying to do here."
You can test this right now in your own workflow. Take a piece of writing or an argument you recently developed. Ask an unconfigured AI for feedback using a neutral prompt. Then, switch to your critical configuration and ask the exact same question. Compare the two outputs.
Ask yourself: Which feedback was genuinely useful? Which one felt comfortable? Did the most comfortable response actually help you improve your ideas, or did it just make you feel good about mediocre thinking?
When we configure our tools, direct them thoughtfully, and refuse empty validation, we engage AI in a way that serves clarity and truth. That is what it means to practice the Zhēn of AI—using technology to cultivate, rather than compromise, our true nature.
