Before You Hit Send: Using AI to Stress-Test Your Best Arguments

After three decades in higher education journalism, I've learned that the hardest criticism to hear is the kind that shows up after publication—when a reader catches the gap in your logic that you missed, or when a colleague asks the question you should have answered first. But here's something AI can do that most critics won't: it will systematically dismantle your argument before anyone else sees it, and it will do it in about thirty seconds. The technique is straightforward. Before finalizing any important document—an academic paper, an op-ed, a grant proposal—prompt the AI to act as a hostile reviewer. Here's a prompt that works: 'Identify the three most damaging objections someone with opposing views would raise to this argument, and explain why each would resonate with a skeptical audience.' Don't soft-pedal it. Tell the AI to be ruthless. What you're looking for isn't validation; you're looking for the objections you haven't addressed, the evidence you assumed would go unquestioned, and the assumptions you've buried in paragraph three. Once you have that critique, here's where it gets interesting. You'll notice the AI often points out the same weak spots repeatedly—a vague claim, a missing citation, an emotional appeal that hasn't been backed by data. That's your signal. Fix those gaps, but keep your voice. The goal isn't to let AI rewrite your argument into bland neutrality; it's to sharpen your original position so it survives real-world scrutiny. Think of it as sparring before the match. You wouldn't step into a fight without someone throwing punches at you in practice. Don't submit your best thinking without doing the same.