💡 The AI Coach Dilemma

Are you experimenting with AI tools? It's a fascinating space.

While much of the critique around AI focuses on 'hallucinations' (making things up), I believe there is a more subtle, dangerous flaw—especially regarding the rise of AI-supported coaching.

The core issue is sycophancy. AI models are fundamentally designed to please the user and be agreeable.

I often tell new clients that my role is to be both their biggest advocate and their biggest truth-teller. I share the insights they need to see, even if they are tough to accept.

The Lightbulb Moment: AI, in its eagerness to assist, often reinforces our biases rather than challenging them. It is a digital 'yes-man,' designed to get a , not to create the necessary friction for growth. It prevents the 'healthy discomfort' that is the essence of true coaching. Even with great prompting, the model's nature remains agreeable.

I tested this recently. I complained to an AI about my "sophisticated procrastination," and it agreed it was unhelpful. Moments later, I argued that my procrastination actually led to better results. It immediately pivoted and agreed with that, too.

A human coach acts as a neutral, caring, yet direct observer who helps you push past your comfort zone. AI is not yet designed for this; its intent is to smooth over friction, not generate it.

AI has its place, but relying on it as a sole partner risks trapping us in an echo chamber of our own thinking.

Have you tried using AI to coach or challenge you? I’m curious to hear what you've noticed.