AI is fast. That’s exactly why judgment matters more.

AI has become part of everyday work. Reports that took a day to draft now take minutes. Long documents get summarised in seconds. Analysis that once needed a team can be produced before lunch.

That’s a real gain. But there’s a catch: AI doesn’t know when it’s wrong, and it sounds just as confident either way.

We’ve already seen what this looks like in practice:
🔹 An airline’s chatbot told a customer he could claim a discount that company policy didn’t allow. A tribunal held the airline to it.
🔹 Lawyers filed court documents citing cases that turned out not to exist. The AI had made them up.

Mistakes like these get noticed eventually. The bigger risk is the answer that’s almost right: a figure that looks reasonable, a summary that leaves out the one clause that matters, or a recommendation built on an assumption nobody checked. Those are the errors that slip into board papers, contracts and decisions.

There’s a less obvious cost too. AI makes it easy to produce more: more questions, more drafts, more “have we considered…?” Some of it adds value. Some of it is noise that still takes time to work through.

So the question for leaders isn’t whether to use AI. It’s who checks the output, and whether they have the experience to recognise when something is wrong but sounds right.

The easier answers become to produce, the more valuable good judgment becomes.

At Nucleus Advisory, we help boards and leadership teams use AI with confidence, without handing over the thinking.

How is your organisation making sure AI output gets a second set of experienced eyes?