Accepting Constructive Criticism from AI Systems

Most of us treat AI like a digital yes-man. We prompt, it delivers, we pat ourselves on the back. But we're missing AI's superpower: being the critic in your corner who spots what you can't see. **The blind spot problem** Our brains are wired for shortcuts. It's how we survive without analysing every decision from scratch. But when you're crafting strategy or pitching ideas, those same shortcuts become traps. AI? It's trained to be helpful. Which means it cheerfully reinforces whatever blind spot you're stuck in. Result: Polished work that misses the mark. **Turn AI into your toughest critic** Forget "AI, write this for me." Start with "AI, tear this apart." My go-to prompts that actually work: • "What assumptions am I making that could be wrong?" • "If you were my competitor, how would you attack this strategy?" • "What would make my target audience stop reading and move on?" • "What evidence contradicts my approach?" • "If this failed spectacularly, what would be the most likely reason?" Specificity is everything. Don't ask for generic critique. Tell AI exactly whose shoes to step into. **The perspective hack** AI can channel any viewpoint if you feed it context. Testing a new product feature? "You're a stressed parent with 3 minutes to decide if this is worth their time. What frustrates you about this?" Pitching to the board? "You're the CFO who got burned on the last innovation project. What red flags are you scanning for?" Rolling out a new process? "You're the team member who has to implement this daily. What's going to make your job harder?" The more context you give about the person's role, experience, and concerns, the sharper the feedback. **Pre-test everything** We used to launch ideas and pray. Now you can stress-test with AI before anyone sees your work. My process: 1. Brief AI on your exact audience (role, priorities, past experiences, current pressures) 2. Ask targeted questions: "What would bore them?" "What seems risky?" "What's unclear?" 3. Fix the weak spots 4. Then test with humans, starting from a much stronger position It's not replacing human feedback. It's making sure you don't waste their time with obvious misses. **Your homework** Take a current project. Ask AI: "What are three ways this could fail that I haven't considered?" Actually listen to the answer. Fix what's fixable. Then watch your work get sharper. The real power isn't in AI agreeing with you. It's in AI showing you what you're too close to see. Ready to level up your AI game properly? Inventium's GenAI Productivity Upgrade starts October 15. We'll show you how to build AI workflows that actually transform how you work. https://lnkd.in/gfeKDvWb #AI #CriticalThinking #ProductivityHacks

Have you argued with your AI recently? Using it to question my answers, rather than just answer my questions has been a game changer. "I'm about to [make this decision/send this proposal/launch this project]. Please challenge my assumptions. What am I not seeing? What questions could I be asking? What could go spectacularly wrong? Be constructively critical." 🤖 It helps fight confirmation bias - Instead of seeking validation, I'm hunting for blind spots 🤖 It's not scared of hurting my feelings or damaging our relationship 🤖 It's endlessly patient - I can keep pushing back: "But what if...?" "Have you considered...?" Obviously then, you can push back and question it further too, as the golden rule is never accept the first answer! Real example from last month, playing with some new ideas: Me: "If I translated my book to Welsh, what should I consider?" Claude gave me a heap of ideas around funding and grants, distribution, translation, promotion and suitable metaphors I hadn't considered. Most of us are surrounded by people who either can't challenge us (they don't have context) or won't (they don't want conflict). AI fills that gap perfectly. Don't just use it to make things faster. Use it to make things better. Remember: "nibble rather than scoff" - start small, take baby steps. #AIDay #CriticalThinking

AI wants to compliment you, don't fall for that. You cant get better when everyone is trying to tell you how good you are. If you are working with AI for critical feedback (I'm doing that now), consider telling it what you really suck at, and remind it that you are not looking for "encouragement" but you want direct, blunt, and harsh criticism. It seems the more I play with AI across all models, that it is widely trying to appease you by telling you how smart you are and how much you are improving, because that is what the average person wants in feedback. But if you don't want affirmation in feedback but improvement, consider prompts like: We're not friends, you are the toughest but wisest boss anyone has ever had, I already know I am poor at these 5 areas. I'm long winded, I tend to meander in thoughts, I tend to be overly tactical when strategy is needed, and I'm really bad at carrying a common thread through my ideas. I am going to provide you this presentation, and I want you to aggressively highlight where I am making mistakes that will prevent me from my goal. My goal is for people who watch this presentation to walk away with a deeper understanding of "how SEO is changing" and I want them to walk away being able to build a plan to keep their company ahead of the curve. Then intentionally miss parts of this to see which models are able to be best at finding your flaws. Again this is where having spoken for years and having 10,000+ people leaving you feedback (good and bad) helps you over others because you already know where your presentation style doesn't hit, where you make mistakes, etc. Harvest that feedback and ask for more of it.

Here's a simple tip for getting honest feedback from AI assistants: Don't ask "What do you think of my idea?" Instead, ask for "pros & cons of my approach, and some alternatives." I've noticed this consistently in my interactions with Claude, Gemini, and ChatGPT. These models are trained to be "helpful," which often means they validate even questionable ideas rather than risk discouraging you. The number of times Cursor has enthusiastically praised my most dubious code proposals, or ChatGPT has complimented me on an obviously terrible product idea is genuinely worrisome. But when I frame the question as an analysis request, the quality of feedback improves dramatically, because it can tell you your idea stinks without saying your idea stinks. By explicitly inviting critique within a structured format, you bypass the AI's people-pleasing tendencies and get something far more valuable: honest assessment. This won't stop me from asking Cursor if I'm the most brilliant engineer it's ever seen for a little ego boost every once in a while, but it helps to be mindful about what you want from your AI responses 😊

Using AI to review, critique and improve my writing is currently my favourite use case. Here's why ⤵️ Below, I'll share the simple prompting framework that consistently gets me results. Save this template for future use and let me know how well it works for you! Whilst AI-generated writing is often cringe, generic, inaccurate, and easy to spot (em dashes, anyone?), using AI to get comprehensive feedback on your work can help take it to the next level. Furthermore, I strongly believe that we should not outsource our critical thinking and human judgement to AI. The process of writing is ultimately how we organise, formulate, crystallise, and express our most complex and meaningful thoughts. Having said that, I've achieved great results from leveraging AI as a critical reviewer, and it's not always feasible to get a real human expert to review your work. Here's how to do it: ➡️ Ask the model to critically review your work and score it out of 100, across the following 10 dimensions: 1. Structure and flow 2. Accuracy and subject matter expertise 3. Spelling, grammar and punctuation 4. Style and tone 5. Originality, substance and depth 6. Clarity and ease of understanding 7. Quality and consistency of argument 8. Reader engagement 9. References, sources and engagement with wider work 10. Miscellaneous ➡️ Demand comprehensive and actionable feedback across each of the 10 dimensions. ➡️ But also ask for the top 5 most important pieces of feedback, so you stay focused. ➡️ Crucially, ask the model to assume a particular persona which is relevant for the task at hand. This could be something like: "Perform this review work as if you are a leading academic authority on the topic of European copyright law who is an esteemed professor at a top university" ➡️ Finally, avoid model sycophancy by instructing the model to be candid, direct and harsh if necessary. If you want to take this really seriously you can also: ➡️ Use the same prompt, but instruct the model to adopt different personas, such as copyeditor and proof reader, or your harshest critic (if you're feeling dramatic). ➡️ Use the same prompt template, for the same piece of work, with different models, like Claude 4, Gemini 2.5, and o3. ➡️ Leverage few shot learning by providing the model with copies of your best work, if it's something you want to emulate. Some words of caution: Don't use AI to review other people's work, or work you have jointly authored, without their permission. Most importantly, don't use this as a full replacement for your own editing and review process. Although you will get plenty of useful insights and feedback, you will always be the best editor of your own work. However, if a piece of writing or publication is particularly important for you, then using the best tools and technology at your disposal makes sense. I didn't use AI to review this post, because it's not that deep...

One of the biggest unlocks in my personal use of AI has been using it as a devil's advocate. Instead of trying to mentally filter responses to correct for AI's default sycophancy (OpenAI's recent update/rollback shone a light on this, but it's a problem with all AIs), I specifically prompt the AI to challenge my ideas and suggest constructive ways to make them better. I've created specific custom GPTs for this, but you can create this kind of interaction in any chat with the right prompt. Here's a good one: "Act as a strategic challenger. Your job is to stress-test my ideas—find weak points, question assumptions, and highlight blind spots. Keep your tone constructive but direct. Don’t flatter me—sharpen my thinking. Ask tough, thoughtful questions and suggest improvements I may have overlooked. Help me turn good ideas into great ones by pushing for clarity, logic, and impact." Using AI in this way is just one of the tips I talked about today in a LinkedIn Live I did with Linda Zebian at Muck Rack: 5 Essential Skills Every PR Pro Needs in 2025. Over 250 poeple watched live, and a bunch of them asked thoughtful questions, including the extremely timely, "Is using AI cheating?" If you missed it, you can check out the recording (I'll drop the link in the comments), plus I'm giving away the slide deck, which is loaded with prompts, tips, and examples. Enjoy, and please drop a like or comment if you find it useful! Grab the slides here 👉 https://lnkd.in/epfF8e-B