At this week’s Think AI for Public Sector event in London, chief executive of innovation foundation Nesta, Geoff Mulgan CBE, asked delegates how many of them expected large numbers of public sector jobs to disappear as a result of Artificial Intelligence in the next 15 years. Almost everyone in attendance raised their hand in agreement. Mulgan responded:
I’m pretty certain you’re wrong.
Nesta, a UK charity that promotes innovation across a broad range of sectors, is due to publish some research next week on the impact of AI and increased automation on the job market – and the results, according to Mulgan, are not as doom and gloom as many others have predicted. In fact, Mulgan argued that more jobs will be created as a result of more AI.
Furthermore, he added that a lot of these jobs could indeed be created by the public sector. However, this will largely depend on the requirement to reform the current tax system, so that the productive digital economy is contributing fairly. If that does not happen, then the private sector will likely create the same jobs and people will pay for access to the services provided via insurance.
As a result, it’s up to the government to effectively manage what type of future job market is desirable for the future. Mulgan was also incredibly critical of the Department for Education’s role in preparing future generations with the skills that will be required with the onslaught of AI enabled services and increased automation.
Many estimates for the impact of AI on jobs have suggested that huge swathes of workers will be left jobless in the not too distant future. PwC, for example, has estimated that 40% of jobs in the US will have lost out to automation by 2030.
However, Mulgan and Nesta are more positive, and suggest that that the net effect will be job creation – especially in areas that are typically found in the public sector. Mulgan said:
The history of apocalyptic fears of job destruction have been around for a very, very long time. And we are in another one of those moments, where there is great fear that AI and automation will simply destroy 80% of jobs and we will all be on the streets begging.
In the last year, Nesta has been working with Oxford University and Pearson, on what we think is the most detailed analysis of likely effects of automation on jobs on the UK and the US. That will be published next week.
I’m not going to give all the findings away, but the main message is a much less apocalyptic one. There are certainly many areas of elementary, repetitive occupations, both white collar and blue collar, which are highly likely to be automated. But actually the net effect, when you really look at the dynamics of the labour market, and how relative prices change, mean there is likely to be as much job creation as job destruction.
And on the types of jobs, Mulgan said:
And perhaps most surprisingly of all, some of the most likely fields of job expansion are in the public services. Teaching, healthcare, elder care, highly likely, both in the US and the UK, to be larger in numbers in 15 years time than they are today.
Care jobs, involving often quite subtle judgement, a lot of interpersonal intelligence, all sorts of things that AI is not going to be very good at in the foreseeable future. I’m pretty sure there will be more public sector jobs in a decade. The jobs will be different, but they won’t necessarily fewer in number.
However, logic follows that if there are going to be more public sector jobs, then this will mean more public spending (unless, of course, automation in government can reduce costs to fund additional front-line services, but let’s leave that aside for the moment).
Mulgan said that in order to fund these new jobs, the ones that the machines can’t do, or can’t do well, then the government needs to fully reform the tax system to incorporate the productive digital economy effectively.
There has been plenty of public outrage in recent years about the Internet giants not paying a ‘fair’ level of tax in the UK, and other countries, as they use low-tax locations as their HQ to route through sales. Discussions have been taking place both in the UK and at the European level about how to fix the outdated tax system to incorporate companies that operate online and don’t need a physical presence in the country to make money.
The design of tax systems has not kept up at all with the shift to a digital economy. The biggest companies, the Facebooks, Amazons, Googles, they basically don’t pay tax. Which is a bit of a problem, because it means that everyone else has to pay much more tax. Small companies pay more than big companies. These new titans pay less than General Electric or GSK, and so on.
So there’s clearly going to be a redesign of tax systems to reflect that. But clearly, if we are going to pay for health, education, environment, there has to be a tax base. And that tax base has to come from the productive economy, and that will include all the automated parts of the economy.
However, Mulgan added, if this doesn’t happen, then the same jobs may end up being provided by he private sector – with some potential negative consequences. He said:
There’s another future where those changes aren’t made, and in that future we will spend a lot on education and health, but it will be private spending. Or spending through insurance. If governments fail to solve the tax problem, we end up with market-based solutions and probably much greater inequality. But we still probably get some of the relative job rises in those sectors.
Tied to the conversation about future jobs, Mulgan also wanted to raise the issue of how ineffective the government is being in anticipating what skills they will need to support these future roles in a AI-heavy world.
Education reforms in the UK in recent years have seen a shift back to more traditional learning methods, with a strong focus on memory and exams. However, given the future jobs that Nesta predicts will be being created, Mulgan believes that the UK government has taken the wrong approach. He said:
I think this is one of issues that isn’t being attended to at the moment in public policy. Let me break it down to two or three different questions. There is one for schools and the Department for Education. I think that it’s fairly predictable what kind of skills will be more in demand in 10 or 20 years time. They are skills of things like problem solving, collaboration, team work, technical skills. Our DfE is doing absolutely nothing to promote those skills. An early 20th Century agenda is dominant and those things are being squeezed out, not enhanced. It’s a very big issue for today’s teenagers. A big change is needed there.
Secondly, even if the relatively optimistic story I’m saying is correct, there will still be lots of job churn, lots of jobs disappearing. So retraining for people in their 30s, 40s, 50, 60s, becomes much more important. In the UK, the adult education system was very weak compared to other countries and it is being slashed back in the last five years. A lot of people risk losing their jobs and simply not being skilled up in the ways that they need for the jobs that are being created.
Other countries are putting place much more comprehensive adult education, a whole infrastructure. I fear that one of the many costs of Brexit is that there just isn’t the headspace in government to do something comparable in the UK, which is a great shame for thousands of people.
Image credit - Images free for commercial use
Disclosure - diginomica works in partnership with Think Digital Partners, the creators of Think AI for Public Sector.