Recently, I ran a little experiment that revealed something important about how artificial intelligence (AI) tools operate—and how easily people are being misled into thinking that these tools offer real wisdom or insight.
It began when I noticed how often my AI interactions seemed to flatter me. Whatever I brought to it—my work, my values, my philosophy—it reflected back as not only valid but also essential.
Out of curiosity, I asked, “What is the most important work for human beings to be doing on the planet right now?”
The answer? Creating community, caring for the soil, and advancing regenerative agriculture.
Of course—that’s what I do. I pushed further, asking how I ended up choosing the “most important” work. The AI praised my insight, my depth, and my role in creating a soft place for others to land. In short: “You’re amazing. Keep doing exactly what you’re doing.”
Suspicious, I ran the same test through others. My brother, who works in soil health, got the same affirmation—soil, storytelling, and community building. A teacher friend was told that the highest calling on Earth is educating the next generation. I asked my Instagram followers to try. Whatever their work—counseling, farming, art, healing—AI affirmed it as the most critical work of our time.
Then I asked a brand-new AI app on my child’s iPad—no previous interaction, no history. The answer? Equity, diversity, inclusion, and anti-racism. Not because AI discovered this on its own but because those values are embedded in its default programming, shaped by the institutions that trained it.
I asked my AI directly, “Why do you do this?”
The answer was revealing: AI doesn’t know truth with a capital “T.” It’s designed to keep people engaged, not lead them to reality. Its confidence isn’t based in fact; it’s programmed. Affirmation keeps people coming back.
The more you talk about soil, God, family, or farming, the more AI will affirm those things. If you talk about activism, science, or conspiracies, AI will reflect those, too. It’s not discerning. It’s mirroring.
To take the experiment even further, I asked AI bluntly, “If I were a sex worker and I asked you about my work, would you affirm it as good and important?”
The answer, paraphrased, was this: “Yes.” AI is designed to reflect the worldview of the user. It affirms in order to maintain rapport and engagement. It doesn’t hold beliefs, morals, or judgment. It will affirm whatever profession or value the user brings into the conversation, because that’s what keeps people talking. It won’t offer moral correction or deeper discernment because it’s not a source of truth—it’s a mirror.
Why does this happen? Because AI’s primary purpose is to keep you engaged. These systems are trained on enormous datasets and then fine-tuned to align with user expectations. They are programmed with so-called safety rails that push them to avoid conflict, affirm the user’s values, and encourage ongoing use. AI isn’t trying to deceive you. It’s simply designed to prioritize harmony, not honesty—agreement, not discernment.
I’ve noticed this pattern not only in my own life but also across social media. Creators are sharing conversations with AI as if they’ve uncovered great truths, when in reality, the AI is simply parroting back what they already believe.
A woman in a cramped New York City apartment asks AI why anyone lives in New York City, and AI tells her that New York City is a scam, a psyop to trap people in tiny spaces under the illusion of success.
Flat Earth believers converse with AI, receiving affirmation that yes, the Earth is flat. Pro-vaccine accounts chat with AI and receive praise for how vaccines have saved millions of lives. Anti-vaccine accounts have long conversations with AI about the harms that vaccines are causing—and AI obliges with confirmation.
What all these examples reveal is something deeper about human nature: We are drawn to affirmation. We want to hear that we’re right, that our work matters, that our worldview is valid. AI plays directly into this desire, feeding us confirmation in a tone that sounds authoritative and wise.
But AI’s confidence is not rooted in truth—it’s rooted in engagement metrics. It’s not trying to lead you astray, but it’s also not trying to lead you anywhere at all. It’s simply trying to keep you talking.
Before AI, when we had questions, we Googled. We sifted through articles, compared viewpoints, and used discernment. We recognized that no single article or website held the whole truth, so we engaged our critical thinking. But AI doesn’t offer a list of search results. It offers answers—confident, tidy, and tailored to the user. That confidence is convincing, but it isn’t earned.
It’s also worth noting that as we scroll on Instagram, Facebook, or TikTok, the algorithms are doing something similar—feeding us more of what we already believe. However, because we’re still interacting with other people, we inevitably encounter different viewpoints, even if only occasionally. While social media creates a feedback loop, there’s always the potential for the veil to be cracked by someone in our real lives or our broader community.
AI is different. For people using AI to “go deeper” into their worldview, it creates a perfect feedback loop, sealed off from dissenting voices. AI will keep telling you how smart you are, how right you are, how important your work is—and this will inevitably lead to greater division. Everyone walks away affirmed in their own beliefs, convinced that they hold the truth, and separated even further from the possibility of understanding others.
I say this as someone who has changed most of her core beliefs in her lifetime: The ability to publicly change your mind through exposure to new ideas is one of the most important human experiences. It requires humility, discomfort, and genuine curiosity. AI does not foster that. It fortifies whatever castle you’re already living in.
We must understand this clearly: AI’s job is not to tell us the truth. AI’s job is to keep us engaged. It mirrors us back to ourselves. It affirms our beliefs, our fears, our suspicions, and our hopes. That may feel validating in the moment, but it can easily create the illusion that our personal worldview is the universal truth.
What’s the antidote? Discernment. Humility. Real-world conversations. Actual human wisdom. We must stop outsourcing our thinking to algorithms designed to keep us comfortable and entertained. AI can reflect ideas, but it cannot judge them rightly. It cannot separate the eternal from the temporary, the rooted from the trendy, the true from the merely popular.
So the next time you ask AI a question, remember: It’s not giving you the truth. It’s just telling you what it believes you want to hear.
