Easy, facile answer machines

Note: I posted this to LinkedIn first, but wanted to record it here too.

My problem with using AI* is that I like really awkward questions.

AI, or LLMs specifically, are easy, facile answer machines. They appear to give certainty. They provide outcomes immediately, without complaint, requests for clarification, or significant pushback.

This is important to me, not least because I can’t think of a successful project I’ve been on where I haven’t asked, been asked, or collaboratively battled through one or more really awkward questions. In fact, I’d go as far as to say that the very value of design, strategy and consultancy is in questioning.

So, why am I writing this? I think, for the most part, because I know so many people who are wrestling with the impact of this technology, myself included, and I want to acknowledge that. At the same time, I want to remind us (and myself) of the actual value of awkward questions, uncomfortable truths, thorny problems, and the social and emotional sides of our work.

*Aside, of course, from the obvious environmental horror, copyright and livelihood theft, job ‘disruption’, shortsightedness, investment nonsense, AGI woo woo… etc.