In 2017, a government minister made a startling comment to me: “Why would we train people as truck drivers? Driverless lorries will render the role obsolete.”
Four years later – during the autumn of 2021 - we were in the middle of huge HGV shortage.
The government responded by providing funding to train more HGV drivers rather than wait for driverless lorries. Even in 2024, driverless trucks are still a long way off, perhaps another decade away.
Predicting the Unknown
This example illustrates that the paths for the adoption for new technology are difficult to predict in both pace and substance. And if you are in the skills business, this is a problem.
At the risk of borrowing from the late US Secretary of Defence Donald Rumsfeld, the idea that artificial intelligence (AI) will impact on the labour and skill needs is a Known, Known.
What is a Known, Unknown is how deployment will happen, and the labour market reshape.
Skills providers will recognise this world of known, unknowns.
Go to any employer and ask them what their skills needs will be in 10 years. Their answer will likely be some variation of ‘what we need now’ or ‘we don’t really know’.
The truth is we don’t really know. And, as we approach AI, false certainty about what we need could lead us up the wrong path.
But we can’t do nothing. If we can’t prepare in the specific, we can prepare in the general. We can build the nervous system of an approach to skills that works.
There are six highly interconnected things that I think really matter.
Improved information flows
The first is improved information flows between workers, their employers, skills providers and public authorities.
The best case for this is effective local and sectoral discussion about immediate needs and how they are developing. We are starting to see, in some of the devolved Combined Authorities, industrial and skills plans finally being brought together.
Public policy that enables transition
A second area for action is a clear focus on transition in public policy. AI needs to be at the centre of an industrial strategy which in turn is deeply connected to labour market and skills policy.
Any industrial strategy that works will need people and people transitions at its heart. Systems to support transitions, not just skills funding, should be on our mind.
With AI disrupting the forms of jobs that are created, we can only protect income for workers and efficiency for public services if the pathways are easy to walk down first time around.
Part of a focus on transition is being prepared to have a bigger debate about employer funding of skills.
The well has been poisoned a bit here by the imposition of an ineffective system – the Apprenticeship Levy – for political aims. But as we know from other things like pensions reform, employers will commit their own funding and participate in government programmes if they see a clear ROI (return on investment). Finding that ROI for skills in the world of AI matters.
Any new scheme should build business support using levy money more flexibly to target provision on emerging areas of high need, especially for young people. But we need the levy to support flexible and targeted provision for later career transitioners too.
The answer here is in the capacity and specialism we give our skills system. Better, multi-year funding arrangements for FE and skills providers - where public money crowds-in private money - will be essential.
This capacity will allow partnerships to form that engage people with an understanding of the next two or three steps – even if they and their (potential) employers can’t see ten steps ahead.
We also need to accept that not everyone has an employer who is willing or able to train workers for the AI transition.
The pandemic reminded us that the UK labour market is full of sole traders, self-employed workers and other forms of contractors. They need support too.
With over 7m job changes a year in the UK labour market, the adoption of AI will take place in the context of a highly flexible and relatively dynamic labour market.
That’s why it is time to look again at Individual Learning Accounts.
When we look at the issues of the system twenty years ago, its failings are obvious – too much producer control, lack of advice related to choices, and lack of public visibility of the system itself.
These are all soluble by a sober government taking a long-term view – and would lead to power being put back in learners’ hands.
Supporting the unemployed and inactive
Where we still fall farther short, however, is in the proper engagement of the unemployed and inactive in skills building and skills transitions.
We can achieve a lot by bringing careers advice and Job Centres into the loop too. A bigger step would be for any future government to make Job Centres primarily about being labour exchanges, not benefit payment centres.
It also means allowing workless adults to train and claim, and retrain and claim, so they have the necessary AI skills to fill AI created and affected jobs.
All of these things would help to create and satisfy sufficient demand. But we need qualifications and curricula that are fit for purpose. The game isn’t about bums-on-seats. We need to involve providers, learners and employers in a discussion about making curricula AI-adjacent.
A skills system that meets the moment
If we had the networked structures described above and enough of a focus on preparing mastery in the areas we know will be relevant, specialising late as technology changes, means we could have a system that stands a chance of meeting the moment.
It requires a level of openness, localism and market-design that will stretch the capacity of the state. But the alternative doesn’t look like progress at all.
Neil Carberry is Chief Executive of REC - the Recruitment & Employment Confederation