Five Prompts to Turn AI into a Critical Thinking Partner
The Reboot Foundation’s core mission is to elevate critical thinking, reflective thought, and analytical reasoning in a technologically complex world. We believe technology should enhance human intellect, not erode or replace it. The key to remaining a strong critical thinker in the Age of AI lies in how you interact with these powerful tools.
In the emerging economy, the winners won’t simply be those who use AI the most, but those who apply it best. While average users may treat AI as an “answer machine” that automates routine tasks, high-performing thinkers can use it to elevate their own intellect by treating the technology as an active cognitive partner rather than a shortcut.
Gaining this competitive edge requires shifting how you interact with these tools: moving beyond basic queries and using AI as an intellectual sparring partner that sharpens your analytical edge.
Configuring your AI platform to play that role isn’t difficult. What’s required are prompts and instructions that will turn off its sycophantic tendencies and require it to question your logic, challenge your reasoning, and point out your blind spots. When presenting counter-arguments, questioning your assumptions, or playing a relentless devil’s advocate, AI can be an effective critical thinking partner.
Below are a few prompts to get you started.
Asking AI to Play Devil’s Advocate
Classical debate often requires a “devil’s advocate” whose job it is to vigorously oppose a point of view in order to expose its flaws and shortcomings. When working through a complex decision, a devil’s advocate can help you identify counter-arguments that you might otherwise overlook. A devil’s advocate will force you to defend your logic or arguments against a sharp critique.
Because conversational AI models naturally default to agreeable, helpful responses, they rarely push back unless explicitly instructed to do so. Left unchecked, an AI will simply reinforce your initial assumptions. By deliberately assigning it the role of an adversary, you turn the tool into an objective sparring partner—one that surfaces inconvenient facts, spots hidden logical gaps, and helps you stress-test your position before putting it into practice.
This prompt comes from prompt engineer “frobinson47” on the website musebox.io.
Using AI To Pressure Test Your Arguments
While a devil’s advocate will challenge your position, pressure testing centers on the structure of your argument. During a pressure test, your evidence, logic and conclusions will be subjected to extreme scrutiny to see where it breaks down. The idea is to expose and shore up these weaknesses before an opponent or crisis can exploit them.
One way to pressure test an argument is to develop the strongest, most intellectually honest counterargument. In logic and philosophy, this is sometimes called a “steel man.” The idea is that the “steel man” identifies weaknesses in your evidence and reasoning, giving you an opportunity to refine your claims and build an airtight position.
This prompt have an AI pressure test your thinking comes from prompt engineer “llcoollevo.”
Using AI to Weed Out Confirmation Bias
As mentioned above, the human brain comes pre-wired with dozens of cognitive biases that help us make quick work of mundane, ordinary decisions. The trouble arises when we default to these short cuts when deeper, more reflective thought is needed.
One of the most common cognitive biases is the confirmation bias, which occurs when we seek out information or evidence that supports our pre-existing ideas and we downplay or ignore evidence that discredits or disputes those ideas. For example, a strong supporter of gun rights might dismiss a report that links the availability of firearms to higher crime rates. Or, he might opt to get his news only from pundits that support the Second Amendment, closing him off to points of view that would challenge his beliefs.
AI can exploit the confirmation bias in its users. Sometimes called the “prompt bias,” this is when users give AI leading or premise-laden prompts that all but force the machine to validate your thinking. For example, asking AI, “Why is X bad for the economy?” will likely result in output that validates the idea that X is, in fact, bad economically without presenting a counterargument.
To defeat this, users must explicitly instruct the AI to seek out counter evidence and to flag prompts that ask leading questions or contain assumptions. In this way users can protect themselves from unwittingly falling prey to the confirmation bias.
Using AI to Avoid the Framing Effect
The framing effect occurs when a decision isn’t based on the facts themselves, but rather how those facts are presented. For example, if an operation is described to a patient as having a 80% success rate, that patient may feel more confident about the procedure than if they were told it had a 20% mortality rate.
When relying on how information is framed for us, we often fail to give it the scrutiny it deserves and as a result our problem solving skills are sidelined. We can use AI to reframe a problem so that we must consider it from a different point of view. This can help us be sure we are responding to the reality of a situation, rather than how it was presented to us.
When users treat AI like an easy answer button, they avoid the work and effort required to build their critical thinking skills. By controlling the way AI responds to queries, you can ensure that your brain does most of the heavy mental lifting. Using the five critical thinking prompts outlined here can help you leverage AI to strengthen your own thinking skills that can then be used in any scenario, on any platform, whether AI is present or not.