AI produces immediate and visible benefits. Work gets completed faster. Information becomes easier to access. Ideas can be developed and expressed remarkably quickly.
Those benefits are real.
But research and real-world cases are also pointing towards quieter changes:
This work has been informed by emerging research into critical thinking and AI use, reported cases of unchecked AI-generated work and growing concern about knowledge and understanding becoming weaker inside organisations.
Selected sources include research from Microsoft and Carnegie Mellon, AP reporting on the Deloitte Australia case and Harvard Business Review’s work on organisational knowledge decay.
I created the AI Use Modes Model™ as a practical way to examine how people use AI and the behaviours forming around that use.
People generally use AI for three broad purposes.
The Human Loop™ turns the methodology into practical action.
It introduces deliberate moments of human involvement before, during and after AI-supported work.
There’s a moment every driver knows. Your satnav tells you to turn left, you see a river and for a split second you trust the voice more than your eyes. That moment of blind trust is funny on a country lane. It’s less funny when millions of us are handing off decisions to large language models at work. At the start of 2026 this wasn’t just a metaphor: an Amazon delivery driver followed her GPS onto The Broomway, a 600‑year‑old footpath across the Essex mudflats...
Last year, a report prepared by Deloitte for the Australian government came under scrutiny after researchers discovered references to academic papers that didn't exist...
AI makes your work sound confident, but can you explain the thinking behind it? Explore how effortless polish triggers a new kind of imposter syndrome....