When AI Starts Giving Us Advice

For most of my life, advice came from people. You asked a colleague who had encountered the problem before, called a friend whose judgement you trusted, consulted an expert or simply made the best decision you could with what you knew. Artificial intelligence introduces something quite different: advice that is immediate, inexpensive and available whenever we want it.

That accessibility is one of AI's great strengths. Expertise has never been distributed equally, and neither has access to people willing to share it. Someone starting a business, preparing for an interview or trying to understand an unfamiliar subject can now have a useful conversation within seconds. AI cannot give everyone an experienced mentor, but it can open doors to knowledge that once depended heavily on whom you happened to know.

There is also freedom in being able to ask questions privately. We sometimes avoid asking other people things because we worry the question is too basic, too personal or not yet properly formed. With AI, you can explore an idea before you are ready to share it, ask for an explanation five different ways or test an argument without having to defend it. There is real value in having somewhere to think out loud.

I have never believed that good judgement means relying entirely on your own judgement. In our family business, one of the most important decisions we made was to trust people from outside the family to advise us. They brought perspectives we did not have, questioned assumptions that had become familiar and were sometimes able to see the business more clearly precisely because it was not theirs.

I believe that willingness to listen to outsiders is one of the reasons the business is still standing today. Family businesses can become particularly attached to established ways of doing things because history, relationships and commercial decisions are so closely intertwined. An independent voice can ask a question that someone inside the organisation may hesitate to ask, or recognise that something which has worked for decades may not work for the next decade.

That experience makes me receptive to AI rather than suspicious of it. A useful adviser does not need to share your background or arrive at the same conclusions you would. In fact, the value often lies in the opposite. I am particularly interested in using AI to construct the case against an idea, identify assumptions or suggest an interpretation that I might not have reached myself.

But there is an important difference between a human outsider and an artificial one. The advisers I have trusted over the years brought their own experience with them. They had made mistakes, watched businesses succeed and fail, dealt with difficult people and lived with the consequences of their recommendations. Their advice was shaped not only by what they knew, but by what had happened to them.

AI has knowledge of a different kind. It can process an extraordinary range of information and identify connections at a speed no individual could match, but it does not experience the consequences of the choice it recommends. If I make the wrong business decision after consulting AI, I live with the result. The machine does not. That seems to me an important distinction when we begin asking it not merely for information, but for advice.

There is another complication. AI usually understands a situation through the way we present it. If I describe a disagreement, I decide which details to include, how to characterise the other person and what I believe the problem to be. A convincing answer may therefore reinforce assumptions hidden inside the question. Sometimes a good adviser needs to challenge the person asking for advice before addressing the problem itself.

As AI becomes more personalised, it may get better at doing exactly that. A system that understands our habits, previous decisions and recurring mistakes could become a remarkably perceptive sounding board. Yet there is a paradox here: the better AI becomes at knowing us, the easier it may be for us to rely on it. Advice that feels perfectly tailored can be particularly difficult to question.

My experience with outside advisers has taught me that seeking advice and surrendering judgement are very different things. You listen because another perspective can reveal what you cannot see from where you are standing. But eventually somebody has to decide, accept the consequences and take responsibility for what happens next. That responsibility cannot be outsourced simply because the advice becomes more sophisticated.

The next generation may have access to something my generation never had: a collection of advisers that is always present, knows an extraordinary amount and becomes increasingly familiar with the person asking the question. That could lead to better decisions in business and in life. But I suspect the people who benefit most will not be those who follow AI most closely. They will be those who learn how to surround themselves with intelligence — human and artificial — while retaining the independence to decide when not to take its advice.