Why there will ALWAYS be work for humans


There is massive uncertainty on the future impact of artificial intelligence.

Among those who we can consider the ‘experts’ – the most qualified on the planet to judge – there are deep disagreements on the potential for general artificial intelligence, the evolution of work, whether AI is an existential threat to humanity, and almost every other aspect of the impact of AI.

Let us leave aside for now the full scope of the future relationship between humans and machines.

On the subject of work, I have frequently found myself bemused by the many people who appear to believe that machines will before long do all work, leaving nothing for humans to do other than hopefully bask in the leisure we have.

While it is possible that fewer people will be in gainful employment (which is not a given, more on that in another post), I don’t believe we will ever have a world of no human work, for many reasons.

What is ‘work’?

Our views on what work is clearly need to evolve for a changing world.

Some people seem to implicitly define work as anything we don’t want to do.

However I believe work is at the heart of what it is to be human. It is deeply human to strive to achieve results, to get better at what you do, to try to contribute to others, to compete, and to be recognized for what we are good at.

Life’s journey is largely about fulfilling the Delphic maxim to “know thyself”, and we discover ourselves in finding what it is we are best at.

The Japanese concept of ‘ikigai‘, illustrated on the left, captures many facets of this quest.

Certainly what the world needs and what you can be paid for will change substantially as machine capabilities rise. As such there will be a different intersection with what we love and are good at.

What happens when machines are better than people at everything?

The range of tasks at which machines are better than human is rapidly expanding.

Certainly a key landmark will be when all the work that people do not want to do is automated. However where people will continue to want to work is at tasks at which they are good, that stretch them, that create value for others.

I strongly question whether machines will ever be better than the best humans at roles such as business strategy, motivational speaking, psychotherapy, storytelling, grief counselling, or hundreds of other roles that stem from the distinctly human capabilities of expertise, creativity and relationships.

However let’s for argument’s sake say that machines become better than humans at everything.

This absolutely does not mean there will be no work for humans.

Where work might still exist

If work will continue to exist after machines transcend our capabilities, it should fulfil two criteria: people want to do it, and people or organizations are prepared to pay for that work to be done. (This assumes that people have incomes even without work. While this is another set of scenarios to explore, the alternative would be complete social destruction.)

Let’s consider just a few examples.

If you could be treated by a machine or a human psychotherapist, which would you choose? If AI can better determine therapeutic paths, it can be used to advise or support a human?

Personal trainers tell their clients what exercises they should be doing (and can also be supported by data and analytics in this), however their primary role is motivating people to exercise. As I often say, we are unlikely to respond to a machine telling us to do another 10 pushups, but we will if they are a person we like and respect.

In a world of greater leisure, unquestionably entertainment will be a cornerstone of the economy. AI and CGI generated actors may win Oscars for their performances. But audiences will prefer their stars to be real humans that they can truly relate to.

Computers may write better songs and create better art, by some measures. But since music and art are expressions of emotion, and machines do not feel emotions, so their products will be artifices. We will prefer the creations of humans. And to be sure, art of all kind is work, usually very hard work.

There will be entrepreneurs and human-driven ventures forever. Perhaps they will be in competition with AI-managed companies. But even if human entrepreneurs struggle in competing with AI-only ventures (which I greatly doubt), they will still choose to do it, as a challenge, to demonstrate their mettle, to seek to improve the lot of themselves and their families.

As I noted in the ABC feature the AI Race, perhaps we will pay people to have conversations, rather than that be bundled in a social interaction.

Creating the work we want

I continually emphasize that the future of work is for us to create. Nothing is inevitable. Enormous positive possibilities are open to us.

I for one will be working for the rest of my life in some guise, and most of the people I know will want to work, to be the best they can be, to create value for the world for the indefinite future.

Thinking about the work that we both want to do and will have value, potentially even when machines can transcend human capabilities, is critical.

There will always be work for humans, and if we go about it the right way it will be exactly the work we most want to do.

  • David Schatsky

    Thanks for making the effort to address this topic. I also believe that a world without work is highly unlikely. But I arrive at that view by a different path.

    You say that in a world where machines are better than people at all types of work, people will find employment only if the work in question fulfills two criteria: people want to do it, and people or organizations are prepared to pay for that work to be done. But people have always done work the don’t especially want to do–in exchange for something they do want. There is no reason this will change in a society suffused with automation. And automation is never free. Designing, building, deploying, and maintaining machines costs money.

    Even in a highly automated world, only work that someone is willing to pay for gets done. That is why, according to one estimate, 46% of the world people get paid to do today could be automated by currently available technologies–but isn’t.

    • Thanks David, good points. You’re probably right that if remunerative work exists it won’t always be done for love. But if machine capabilities are that good, they will have taken all the work people don’t want to do, the markets will be mainly for human-centered work – which far more people will be happy to do, even if they wouldn’t choose to do it for nothing.

  • Another very important example: teachers. Teaching requires flexibility, empathy, problem-solving, and a human connection that machines will likely never be able to replicate.

    • I agree Denis, noting that we will see a blend of recorded/ remote teaching, machine-learning personalized teaching, and in-person teaching. The first two will be immensely valuable, but are absolutely not a substitute for someone who can truly understand a person, and most importantly motivate them to learn and grow. Learning is, of course, at the heart of our future.