My favourite AI prompt doesn't write for me. It helps me think more clearly about what I've written. A simple way to use AI as a mirror and sparring partner, while keeping my own judgement in the driver's seat.

AI makes it easy to produce convincing answers, but easy isn't always better. I've started deliberately adding friction to my AI workflows through time, effort and human feedback. Not to slow myself down, but to create space for clearer thinking and better judgement.

Saying “please” and “thank you” to AI costs energy, water and money. But what might it cost us to stop? The Politeness Paradox raises an unexpected question: when we change how we treat machines, are we also changing how we treat each other?

It’s not just overtrust, and it’s not just a hallucination. It feels like a kind of cognitive sleight of hand. The Believability Effect.

AI can mislead us without intending to. But here's the uncomfortable part: we help it happen. From projecting human qualities onto chatbots to revealing what we want to hear, our own habits make AI easier to believe and harder to question.

Could ChatGPT replace a friend on your morning commute? I took ChatGPT for a walk and Gemini out for coffee to find out. Both were happy to chat about anything I wanted. But something was missing, and it wasn't just the occasional made-up fact.

Where I explore the appeal and risks of chatbot companionship.

Poking at some uncomfortable questions about human-like AI.