Why Humans Still Matter in the Age of Intelligent Machines

There is a lot of talk at the moment about what AI is going to replace, and I find the whole conversation slightly backward because we seem obsessed with asking what AI can do that humans currently do, rather than asking what it can never really do and why those differences are actually what could make the relationship between humans and AI so powerful.

For all the extraordinary things AI can do, there are some things it simply cannot replicate in the way a human being experiences them. It can analyze emotion, understand language, recognize patterns and produce something that sounds incredibly sensitive, but it doesn’t actually feel anything. It doesn’t know what it is like to walk into a room and instantly sense that something is wrong before anybody has said a word, or to meet somebody for five minutes and have an inexplicable feeling that you should trust them, or perhaps that you absolutely shouldn’t. 

It can understand the language of emotion, but there is a very important difference between understanding an emotion and actually having one.

The same applies to instinct, which I think is one of the most underrated qualities in entrepreneurship. Some of the biggest decisions people make are not based entirely on data, because sometimes you look at something and you simply know that it is right, even though you cannot necessarily explain why. 

You cannot always put instinct into a spreadsheet or build a model around it, because instinct is often the accumulated result of everything you have experienced, every mistake you have made, every success you have had, every person you have met and every lesson you have learned, including many of the lessons you didn’t even realize you were learning at the time.

AI doesn’t have that life experience. It has information, an almost absurd amount of information, and it can process that information at a speed and scale that is extraordinary, but it hasn’t actually lived any of it. It hasn’t made a terrible decision at 28 and spent the next ten years remembering exactly why it was terrible.

It hasn’t walked into a meeting completely unprepared and somehow managed to turn it into the best deal of its career. 

It hasn’t had somebody tell it that an idea is impossible and then become slightly obsessed with proving them wrong, partly because it believes in the idea and partly because it now really wants to prove that person wrong.

I don’t want AI to be me

All that is exactly why I think AI can be such an extraordinary assistant.

I don’t want it to be me. I want it to be the perfect counterpart to me because the very things that make AI different from us are precisely what make it so useful. I want something sitting there that doesn’t get tired, doesn’t need a coffee, doesn’t have to go home, doesn’t get offended when I say, “No, that’s rubbish, start again,” and doesn’t start wondering whether I have any idea what I am actually doing. I want to be able to throw an idea at it at midnight, completely change my mind twelve minutes later, ask it to approach the whole thing from another direction, decide that direction is terrible, go back to the original idea and then suddenly come up with something completely different, and have it still sitting there ready to help.

It doesn’t moan, it doesn’t sulk, it doesn’t complain that we’ve already done three versions, it doesn’t need a weekend, it doesn’t need a holiday and it doesn’t decide that because somebody else got the better project, it is going to quietly resent me for the next six months. It doesn’t need feeding, it doesn’t need sleeping and, importantly, it doesn’t cheat.

Well, not very often.

AI is occasionally an idiot

It is also, occasionally, an absolute idiot, which I actually find quite reassuring.

You can have a brilliant conversation with AI, get to a genuinely interesting thought, and then suddenly it will confidently tell you something that is completely wrong, misunderstand something you’ve explained three times or produce something that is technically perfect but has absolutely no idea why anybody should care about it. The funny thing is, this makes it considerably more human than some people who are currently predicting the end of humanity, because humans get things spectacularly wrong, too.

The difference is that humans can have a terrible idea and then, five minutes later, have a brilliant one. We can become irrationally passionate about something that nobody else understands, we can take a risk when all the available evidence tells us not to, and we can sometimes be arrogant enough to believe that we can do something everybody else thinks is ridiculous. That arrogance, in the right person and in the right circumstances, can be an incredibly powerful thing because, without a certain amount of irrational confidence, a huge number of businesses, inventions, films, products and ideas would probably never have existed.

And there is some pretty compelling evidence suggesting we are not talking about a distant future here. Stanford’s 2025 AI Index found that 78% of organizations surveyed reported using AI in 2024, up from 55% in 2023, while the proportion using generative AI in at least one business function more than doubled, from 33% to 71%. McKinsey’s research has also found that 64% of respondents say AI is enabling innovation in their organizations, although only 39% report an impact on EBIT at the enterprise level. I think this is an important distinction, because adopting AI and actually knowing how to use it properly are clearly two very different things.

That is the bit I don’t want us to lose.

I don’t want a future where humans start trying to think like machines simply because machines are better at processing information. I want the opposite. I want a future where humans can become even more human because machines take care of more of the things they’re good at.

Steve Jobs put it rather brilliantly when he said, “Technology is nothing. What’s important is that you have a faith in people,” arguing that if you give smart people good tools, they can do wonderful things with them. That was said in 1994, long before anybody was talking about ChatGPT or generative AI, but I actually think it is one of the most relevant observations you can make about where we are heading now, because the technology may be extraordinary, but it is still the person using the technology who decides what is worth creating.

Let’s let humans decide

Let AI do the research. Let it organize the information, find the patterns, challenge the thinking, generate the alternatives, and help turn an idea into something tangible much faster than we could have done before, but let the human decide whether any of it is actually worth doing.

Because creativity isn’t simply the ability to produce ideas. It is the ability to care about an idea enough to pursue it when other people don’t understand it, to recognize something special when it appears in front of you and, perhaps most importantly, to have the courage to back yourself when there isn’t enough evidence yet to prove that you are right.

AI can give you a thousand possibilities, but somebody still has to choose one. It can analyze the market, but somebody still has to walk into the room. It can write the pitch, but somebody still has to stand there and believe in it. It can tell you what the data says, but there will always be moments when an entrepreneur looks at the data and says, “I know what it says, but I don’t believe it.”

And sometimes that person will be completely wrong. But sometimes they will be right, and those are the moments that change everything.

That is why I don’t see AI as the end of human creativity. I see it as potentially the greatest creative partner humans have ever had, because for the first time we can have something sitting alongside us that can absorb an extraordinary amount of information, work at extraordinary speed, never get tired and never get bored, while the human being remains responsible for the things that make the work worth doing in the first place.

The human brings the soul, the instinct, the emotion, the ambition, the experience, the irrational confidence, the humour and, occasionally, the stupidity. AI brings the patience, the memory, the processing power, the research, the speed and the extraordinary willingness to have another go without ever saying, “For God’s sake, we’ve been through this already.”

And maybe that is actually the perfect relationship.

One of us needs sleep, and one of us doesn’t. One of us occasionally has absolutely no idea what we’re doing, and the other one can produce a remarkably convincing answer even when it has absolutely no idea what it is doing either. One of us has instincts, emotions, and a lifetime of experiences that cannot be reduced to data, while the other has access to an enormous amount of knowledge and can help turn those instincts into something real.

That sounds less like a battle between humans and AI to me and much more like the beginning of a very happy marriage.

Provided, of course, the human remembers who is supposed to be in charge.