Denis McDonough used to go on night walks with President Obama at the end of the work day. In the course of one stroll, Obama told him about a technology which had struck him during a trip to San Francisco: A new device which could compare one X-ray image against millions of others. “It could be more precise and come to conclusions faster and, therefore, would be cheaper than if a doctor examined it,” says a man who, as Obama’s White House Chief of Staff, used to be responsible for 4,000 staffers. “The President said to me that being a radiologist used to be a good job, but that maybe it was going to disappear.” That conversation stuck in his mind. When his time at the White House came to an end in 2017, he decided to focus on the topic of work, an issue he considers key for national security. “Our ability to handle the transition and effectively employ artificial intelligence has considerable implications for national security.”

Technology has clear advantages but, at the same time, it could make many jobs redundant. Do you think the fear many workers have of losing their jobs is well-founded?

The fears are well-founded. At the same time, they might not be entirely accurate. However, the point is not to explain away somebody’s fears: The point is to address them. The best way to do that is to ensure that our schools are performing at their best and that people have options to refresh their skills when they get out of formal schooling.

Whose responsibility is it to retrain? Should it be left to companies or be taken into governments’ hands?

We need to reform the entire ecosystem of labour markets, which includes employers, educators, and workers. Our goal should be to bring greater transparency. Companies need to be more clear about the skills they need. Educators have to offer them on the marketplace, so that the worker can learn that skill. Local, state and federal governments also need to be part of this reform.

What skills do think workers should focus on acquiring if they want to cope well with the coming transition?

Obviously, there has to be a baseline fluency with technology, but I think this is the easiest part of the equation. The most important skill is knowing how to learn. In this  economy, you are going to have to constantly retrain yourself. It will require a mental shift. Right know, we believe that you are done once you have gone through four years of  college and that, from then onwards, all further training will be taken care of by your company. That is a mistake. If we do not build realistic options for people to continue to  learn, we are going to fall back as a country.

President Trump’s rhetoric on reviving the coal industry might have encouraged workers’ hopes that there is still a future in their jobs. How do you make sense of this?

The important lesson here is that people derive a lot of meaning, worth and dignity from their work. Secondly, I think what we see in in the US is not a reflection of a hope of  any particular job coming back, but a manifestation of the unease that there are no alternatives. As policymakers, we should enunciate a clear set of alternatives. The  change that we have witnessed in the coal industry is a reflection of technology and the price of natural gas. President Trump, instead of looking into job alternatives for the  future, is painting a picture that looks backwards.

Can his discourse be an obstacle to the new skills agenda?

It could be, to a certain degree, an obstacle. But there are much bigger ones: the fact that the price of college is growing at astronomical rates, or the fact we have not removed the stigma from certain jobs. That is something President Trump could do a quite good job of changing. The way he shows such pride in miners and factory workers sends a very positive signal. I just wish it were on jobs those aspirations could be fulfilled in, rather than in sectors which, because of major technology developments, are becoming less labour-intensive.

Is there any job that robots will never be able to replace?

If you look at the past, the kind of tasks that got replaced were repeatable, less cognitive and less sociable ones. Machines do certain things extraordinarily well and efficiently. However, the things they do poorly they do very poorly. We should move towards a future where robots are not replacing humans, but teaming up with humans. In that scenario, I think you will see great productivity gains and, as a result, great wage gains, provided we make the investments we need for retraining and reskilling.