#11 - The skill that stops AI from doing your thinking for you
Because outsourcing your brain isn’t a strategy for long term success.
I’m writing this article with some irony, given that my first action when I sat down at my desk to write this on Wednesday was to open up a new dialogue with my bessie mate Chatty, and ask him for some ideas. Luckily, I caught myself just in time before he outputted some slop for this intro.
If you spend any sort of serious time on Linkedin, work in the tech industry, or both (!) you can’t escape AI. If it’s not social media posts generated by AI, it’s posts criticising the use of AI to generate said posts, which themselves may well be generated by AI! We’re entering the era of ouroboros: the snake consuming its own tail.
Many will claim there are ‘tells’, little linguistic quirks that give away when someone has used AI. Behold — the poor maligned em dash. Once a favourite of esteemed novelists like Austen, Bronte, Joyce et al (on whose work these models have been trained!), its overuse now seems to have turned everyone into Hercule Poirot, hunting for clues in punctuation and grammar. Now here’s the thing: people don’t even necessarily talk like this, let alone write like it. It’s not authentic expression; it’s statistical mimicry (did you see what I did there?)
What I’m building to, and the whole point of this post, is that we need to engage our critical thinking now more than ever. This is the second skill I called out in a previous post deep-diving into why (and how) product leaders need to build their creativity skills. Because whilst we think we can spot the obvious AI-generated content out there, it’s going to become increasingly harder as the technology gets better. So we probably need to start treating it more seriously and apply the same intellectual rigour as we would to a human-generated piece of communication.
And this post is going to give you three techniques to do just that.
Technique #1: A heuristic for critical evaluation
This one comes from the final year of my psychology degree, where I had to critically discuss and/or evaluate statements like:
Minds exist inside cultures, but also cultures exist inside minds;
Happiness is not just a personal experience, but a tool for governance, a measure of social progress, and an objective for social, public and economic policy;
When using psychoanalytic concepts in social psychology, we need to be mindful that they emerged from clinical practice;
Not to mention countless journal articles for my final-year literature review.
The OU is absolutely superb at breaking down how to learn new skills. They taught a heuristic for critical thinking based on looking at a topic from three perspectives: within, between, and meta.
Here’s how it works, using a simple example: buying a car.
Within critical evaluation means examining something on its individual merits. In the car example, that might mean looking at a single model and assessing its fuel efficiency, safety features, and how comfortable it is to drive, all without comparing it to anything else.
Between critical evaluation means comparing two or more things. In the car example, you’d compare, say, a VW Golf with a Honda Civic, weighing up pros and cons like price, reliability, or running costs. Or you might compare petrol versus diesel engines.
Meta critical evaluation means stepping back to think about broader implications beyond the immediate choice. For cars, that might include environmental impact, urban planning considerations, or how increased car ownership affects public health and road safety over time.
This simple within-between-meta lens can be applied to anything: an argument, a piece of research, a business decision, or even evaluating an AI-generated text. It’s a powerful way to dig beneath the surface and think more deeply.
Technique #2: PROMPT criteria
Sticking with what I’ve learned at the OU a little longer (that degree is just paying for itself now!), I also learned when it comes to journal articles it is worth considering the PROMPT criteria:
Provenance: where does this information come from, and is the source credible?
Relevance: is this information directly useful for the question or problem at hand?
Objectivity: is the content unbiased, balanced, and free from hidden agendas? (Top tip: follow the money to check for bias).
Method: how was this information or answer produced, and can that process be trusted?
Presentation: how clearly, accurately, and appropriately is the information communicated?
Timeliness: is the information current and up to date for the context in which it’s used?
I looked on the BBC website just now to find a contemporary news article as an example. The top headline is UK inflation at highest for almost a year and a half. Applying PROMPT we could assess it like this:
Provenance: The BBC is a reputable national and international broadcaster, founded in 1922 — not exactly the new kid on the block.
Relevance: Many people in the UK will be concerned about inflation, particularly if they have significant financial commitments like mortgages, savings, or pensions, or are concerned about the rising cost of living.
Objectivity: The BBC has strict rules about impartiality, especially as it’s funded by UK citizens through the TV licence fee — so there’s a high likelihood the coverage is balanced.
Method: The article includes interviews with economists from the Bank of England, KPMG, and the ONS.
Presentation: The narrative is easy to follow, uses graphs to illustrate changes in inflation over time, and avoids jargon when explaining key ideas.
Timeliness: The article references June’s inflation figures, and given today is 16 July, this is very current.
The same thinking applies when working with AI. Just because an answer sounds polished doesn’t mean it’s trustworthy — there are plenty of examples where AI has hallucinated, and Grok in particular is infamous for spouting biased answers until it’s challenged. Running outputs (including their sources) quickly through PROMPT is one way to make sure we’re not outsourcing our judgement along with our writing.
Technique #3: The Six Thinking Hats
Now this is a technique I really love. I first came across it more than 15 years ago (pre-Product career in fact!) when I was working in social housing on transformation projects. Austerity called for some serious critical — and innovative! — thinking.
Allow me to introduce you to Edward de Bono. He was a Maltese physician, psychologist, and author best known for his work on creative and lateral thinking. He wrote over 60 books on how to improve thinking skills, including the well-known Six Thinking Hats method, which helps individuals and teams look at problems from different perspectives to encourage clearer, more innovative decision-making.
The Six Thinking Hats and their respective focuses (foci?) are:
⚪️ White Hat – Focuses on facts, data, and information. “What do we know? What do we need to find out?”
🔴 Red Hat – Represents feelings, emotions, and intuition. “How do I feel about this?”
⚫️ Black Hat – Critical thinking, caution, and identifying risks. “What could go wrong?”
🟡 Yellow Hat – Optimism, benefits, and positive thinking. “What are the advantages?”
🟢 Green Hat – Creativity, new ideas, and alternatives. “What other possibilities exist?”
🔵 Blue Hat – Organising the thinking process, managing which hats to use. “What’s the next step? How should we proceed?”
What I particularly like about this technique is it’s also a pretty solid counter to group-think. Whilst harmony and early convergence on ideas and decisions can feel nice, it can also be incredibly dangerous (see: Groupthink and the space shuttle challenger accident: Toward a quantitative case analysis, by Esser and Lindoerfer, 1989). The Six Thinking Hats explore multiple perspectives and help define the boundaries of what is and isn’t known. It’s a systematic way to surface those known and unknown “knowns” — before they come back to bite you.
This technique is also invaluable when it comes to AI. For example, if your boss comes to you insisting that “we need to add AI to the product,” you could use the Six Thinking Hats to explore the idea from multiple angles — or at the very least, ask some probing questions to check their thinking:
What data would we need (White Hat)?
How do different teams feel about its implementation (Red Hat)?
What risks or downsides could it create for our users (Black Hat)?
What benefits or competitive advantages might it bring (Yellow Hat)?
Are there creative ways to integrate AI beyond the obvious applications (Green Hat)?
And finally, how will we structure the conversation and decide on next steps (Blue Hat)?
So that’s it — three techniques to keep your critical thinking sharp in a world increasingly shaped by AI. Whether you’re analysing information on its own merits, putting AI outputs through the PROMPT checklist, or using the Six Thinking Hats to examine new ideas from every angle, the goal is the same: don’t let AI do all the thinking for you. The tools might be getting smarter, but our responsibility to think critically, ask good questions, and stay curious has never been more important.
Thanks for reading. There’s one final article to come in this series, looking at decision making (I’ve got a LOT to say on that one!). For now, think about what you’ve got coming up where you really need to engage your critical thinking, and consider using one of these techniques. Let me know how you get on — and remember, your brain is still your best asset.
— Caroline
When you know how to think critically — but your boss wants shortcuts — how do you push back?
Work with me 1:1 to build the confidence to reframe the conversation and influence decisions.
Or Invite me to your event or podcast and help your audience see things differently.