AI’s default mode is agreeable — it tends to affirm whatever position you present. That’s comfortable and useless for thinking clearly. Learning how to prompt AI to argue both sides of a topic turns it into a sparring partner that sharpens your thinking instead of just validating it.
Why Both-Sides Prompting Matters
We all have confirmation bias — we seek information that supports what we already believe. AI trained to be helpful amplifies this: present your view, get supporting arguments, feel smarter, learn nothing. Deliberately prompting for both sides is the antidote. It’s how you stress-test opinions before they cost you money, reputation, or bad decisions.
Technique 1: The Balanced Brief
“Present the strongest arguments FOR and AGAINST [position/topic]. Give each side its best, most charitable case — no strawmen. Then tell me which side you find more convincing and why.”
The key phrase is “strongest arguments” and “no strawmen” — without this, the AI gives the favored side steel and the other side straw. (A “strawman” is a weak version of an argument, easy to knock down. Steel-manning — presenting the strongest version of the opposing view — is the intellectually honest move.)
Technique 2: Devil’s Advocate Mode
“I believe [your position]. Act as a smart, well-informed critic. Attack my position with the strongest counterarguments. Don’t hold back to spare my feelings — I want the real objections.”
The “don’t hold back” matters — default AI softens criticism. Explicitly licensing bluntness produces genuinely challenging pushback. Follow up with: “Now defend my original position against those objections” to see which objections survive.
Technique 3: Sequential Advocacy
For the deepest analysis, run both sides separately:
- “Argue FOR [position] as convincingly as you can.”
- In a new chat (or after): “Argue AGAINST [position] as convincingly as you can.”
- “Here are both arguments [paste]. Which is stronger, and what would change your assessment?”
Separate runs prevent the AI from hedging or false-balancing — each side gets full-throated advocacy. The synthesis step then judges the real contest.
Technique 4: Perspective Rotation
“Analyze [issue] from three perspectives: an economist focused on efficiency, an ethicist focused on fairness, and a practitioner focused on what actually works. Then synthesize.”
Different lenses surface different considerations. The economist sees incentives; the ethicist sees justice; the practitioner sees friction. Most real-world issues need all three.
Avoiding False Balance
Important nuance: “both sides” doesn’t mean equal weight for everything. On topics with settled evidence (vaccines, climate change), demanding 50/50 balance manufactures doubt. Better prompt: “Present the mainstream scientific consensus and the main objections to it, with a sense of the weight of evidence for each.” Both sides presented; evidence weighted honestly.
Both-Sides Prompting for Specific Decisions
- Career moves: “Make the strongest case for taking this job, then the strongest case against. Weight: which risks are reversible vs. irreversible?” Reversible/irreversible is the key lens for career decisions.
- Purchases: “Argue for buying [X] and against. Include total cost of ownership, not just price.” Counters marketing-influenced thinking.
- Business strategy: “Present the bull case and bear case for [strategy], with the key assumption each side depends on. Which assumption is more likely true?” Assumption-hunting is the real value.
- Relationships: “Help me see this conflict from the other person’s perspective as charitably as possible. Then from mine. Where’s the misunderstanding?” (AI as empathy amplifier — genuinely useful.)
- Investments: “Bull case, bear case, and base case for [investment]. What would change your assessment?” Never invest based on AI alone — but use it to check your blind spots.
The “Strongest Counterargument” Habit
Beyond formal debates, build this micro-habit: for any opinion you hold confidently, occasionally ask:
“What’s the strongest counterargument to [my view]? Steel-man it — give the opposition its best case.”
Five minutes, occasional use, profound effects over time. You’ll notice your own views getting more nuanced, your arguments getting stronger (they’ve survived contact with opposition), and your ability to persuade improving (you understand the other side’s best case). It’s intellectual calisthenics — and like physical exercise, the benefits compound invisibly until they’re obvious.
Teaching Critical Thinking With Both-Sides AI
For educators and parents: both-sides prompting is a phenomenal teaching tool. “The AI argued X — find the flaw in its argument.” “Here are both sides; which evidence is stronger and why?” Students learn to evaluate arguments, not just consume them. In an era of AI-generated persuasion, the ability to steel-man, spot weak arguments, and weight evidence is civic infrastructure. Teach it deliberately — it’s among the most transferable skills there is.
Strong steel-manning requires structured thinking — step-by-step reasoning prompts force the AI to build each side’s case carefully instead of gesturing at it. And when you find an advocacy prompt that works, getting consistent results from the same prompt ensures the technique holds up across topics.
For the history and academic framing of this field, Wikipedia’s article on prompt engineering is a solid starting point.
Bottom Line
How to prompt AI to argue both sides: demand the strongest version of each position (steel-man, no strawmen), use devil’s-advocate mode for your own beliefs, run sequential advocacy for depth, rotate perspectives for breadth — and weight by evidence, not false balance. An AI that only agrees with you is a mirror; an AI that challenges you is a tool.
