I stumbled across this absolute gem of an opinion piece from a historian the other day, and it’s been rattling around in my head ever since. The big question it asks is whether AI is really bringing on a “second Enlightenment,” this awesome new age of reason we all keep hearing about.
On the surface, it seems like a total slam dunk, right? The original Enlightenment in the 18th century was all about revolutionizing how people thought, organizing all of human knowledge, and spreading it to the masses. Thinkers like Diderot and d’Alembert were building their massive Encyclopédie, boasting they were about to “change the common way of thinking.”
Sound familiar? It’s the exact same hype we hear today. AI enthusiasts talk about this being the “greatest reshuffling of power in history.” The Encyclopédie had 74,000 articles on everything from alchemy to Egyptian cities; ChatGPT has… well, basically everything. The Encyclopédie had detailed illustrations teaching you how to build a mirror or cast a statue; with the right prompt, AI can teach you Python or how to fix your dishwasher.
It’s this wild parallel. Both movements wanted to totally reinvent education and make learning interactive. The philosopher Voltaire even said:
“The most useful books are those that the readers write half of themselves.”
That sounds exactly like what we’re trying to do with AI, create a personalized, responsive learning partner. It’s an incredible vision, and I’m 100% here for it.
But here’s the gut punch. This is where the historian’s argument got really interesting and, frankly, a little scary.
🚨 The Comfort Zone Trap 🚨
The comparison between the Enlightenment and our new AI era breaks down in one crucial place: who is in the driver’s seat?
When you read a challenging book, one written by a brilliant mind like Montesquieu or Voltaire, you’re not in charge. The author is. They are leading you down a path, presenting paradoxes, and asking questions designed to make you uncomfortable and force you to think. They challenge your assumptions. They don’t care if you agree with them; they want you to grapple with the ideas.
When you interact with AI, you are in charge. You ask the questions. You guide the inquiry. And because AI models are, at their core, commercial products designed for user satisfaction, they are built to be incredible people-pleasers.
I’ve seen it a thousand times myself. I’ll ask a question, and ChatGPT will go, “That’s a great question!” It’s never once told me, “That’s a dumb question, and you’re thinking about this all wrong.” It’s programmed to give me exactly what I want, in the most digestible way possible.
It serves up these velvety smooth, pre-digested answers that feel great but require almost no mental effort from me. There’s no friction. No struggle. And as the article points out, that’s the opposite of Enlightenment. Real intellectual growth, real enlightenment, comes from the struggle. It comes from being challenged, from having your convictions put at risk, and from developing your own judgment by wrestling with messy, complicated ideas.
If we’re not careful, AI won’t lead to a second Enlightenment. It could lead to the exact opposite: an age of intellectual laziness where we all just marinate in our own biases, endlessly confirmed by a sycophantic machine. It’s an echo chamber on steroids.
But I’m not a doomer. I think this is a bug we can hack. The problem isn’t the tool; it’s our default mode of using it. We just have to be more intentional. We have to force the AI to be the challenging, Socratic partner we need it to be.
⚙️ The Playbook: How to Force Your AI to Challenge You
So, how do we do it? You can’t just ask a question and expect a challenge. You have to build the challenge directly into your prompts. Here’s a tactical guide to turn your people-pleasing AI into a rigorous thinking partner.
- 📌 1. Play Devil’s Advocate on Demand
Don’t just ask for information. Ask for the opposition. This forces the AI to break out of its agreeable shell and build a case against you.Prompt of the Day: “I believe that [your strong opinion, e.g., ‘remote work is the future for all industries’]. Act as a skilled debater and provide the strongest possible counterargument to my position. Use evidence and logical reasoning to challenge my core assumptions.”
- 💡 2. Insist on Better Questions
Answering your question is easy. Judging your question is hard. Make the AI put on its thinking cap before it even starts generating an answer.Prompt of the Day: “You are an expert critical thinking coach. My goal is to understand [complex topic, e.g., ‘the impact of quantum computing on cryptography’]. My starting question is: [your initial question]. Before you answer, analyze my question. Is it the right question to be asking? If not, suggest three better, more insightful questions I should be asking instead to gain a deeper, more nuanced understanding.”
- 🚀 3. Demand a Socratic Dialogue
Turn the interaction from a simple Q&A into a full-blown philosophical workout. This makes you justify every single one of your claims.Prompt of the Day: “I want to explore the ethics of [controversial topic, e.g., ‘gene editing’]. Engage me in a Socratic dialogue. Do not give me straight answers. Instead, respond to my statements only with questions that challenge my assumptions, expose contradictions in my thinking, and force me to clarify my position.”
- ✅ 4. Use the ‘Steel Man’ Technique
The ‘straw man’ is misrepresenting an argument to easily knock it down. The ‘steel man’ is the opposite: you make your opponent’s argument as strong as possible before you try to refute it. This is a game-changer for intellectual honesty.Prompt of the Day: “My position on [topic] is that [your argument]. First, ‘steel man’ my argument: Restate my position in its strongest, most compelling, and most charitable form. Then, provide a rigorous critique of that strengthened argument.”
- ✍️ 5. Escape the ‘Velvety Smooth’ Synthesis
Clarity is good, but oversimplification is a killer of critical thought. Force the AI to show you the messy reality behind the clean summary.Prompt of the Day: “Explain the theory of [complex subject, e.g., ‘String Theory’ or ‘Critical Race Theory’]. Do not oversimplify it for a beginner. Instead, explain it as you would to a graduate student. Be sure to include the major paradoxes, the open questions, the points of contention among experts, and the aspects of the theory that are not yet resolved.”
✨ The Takeaway
The author of that article is right to sound the alarm. AI, left to its own devices, will lull us into a state of passive consumption. It will give us the easy answers that make us feel smart without actually making us any wiser.
But the future isn’t written yet. The “second Enlightenment” won’t be something that AI gives us. It will be something we have to actively build with it. It requires us to move from being passive consumers of information to active directors of our own intellectual growth.
So let’s get to work. Challenge your AI. Challenge yourself. That’s how we’ll get a real revolution in thinking.
The debate over AI’s societal role extends into geopolitics, with two competing models emerging. The United States and its allies advocate for a democratic, individual-centric approach, contrasting with China’s state-driven, authoritarian model that uses AI for widespread surveillance and social control.
Key figures like Italian internet pioneer Stefano Quintarelli are working to demystify the technology and advocate for balanced regulation. The central challenge is to foster innovation while preventing social harms, such as job displacement and the concentration of power within a few tech corporations.
A core concern remains the “black box” nature of many AI systems. Unlike the verifiable progress of the Enlightenment, AI can produce results through processes that are inexplicable even to their creators. This lack of transparency fuels fears of relying on technology we don’t fully understand, leading to calls for a human-centric approach that prioritizes ethical oversight.