The growing capability of machines has raised the spectre of mass technologically induced unemployment and profound economic disruption. Yet despite the accelerating ability of robots and artificial intelligence, we are not on the cusp of a ‘post-human’ economy.

Indeed, in the absence of policy intervention, the most likely outcome of automation is not mass joblessness but spiralling inequality. If we want the machine age to usher in a future of shared plenty instead, we urgently need to reimagine our economic institutions and practices.

The paradox of plenty

Automation risks creating a ‘paradox of plenty’: society is likely to be far richer overall due to the material abundance generated by machines, but for many individuals and communities, technological change could reinforce inequalities of power and reward as the benefits are narrowly shared.

IPPR’s new report on automation argues that, without policy action, rising inequality is likely because the economic dividends of automation are set to disproportionately flow to the owners of technologies and businesses, and the highly skilled, as income shifts from labour to capital and the labour market polarises between high- and low-skilled jobs.

The trend of a rising share of national income going to capital at the expense of labour is likely to be accelerated by automation and digitalisation. For example, the IPPR analysis shows that the total level of wages associated with jobs with the technical potential to be automated is £290bn per annum – one third of the wage total. If automation leads to lower average wages or working hours, or loss of jobs overall, a significant amount of national income could be transferred from labour to capital.

Even if wages do not decline, in recent decades the trend has been for relative rewards to capital to rise more quickly than those to labour. So this implies that the share of national income going to capital would still increase. As capital is narrowly owned, a rising share of national income going to capital will necessarily increase inequality; whoever owns the robots will own an increasing share of national wealth.

Deepmind, Deliveroo, and ‘digital Taylorism’

At the same time, the growing use of AI and robotics is also likely to further polarise the labour market. On average, low wage jobs have five times the technical potential to be automated compared with high paid jobs. Different regions and sectors are also variably susceptible to automation. London has the highest proportion of jobs assessed as more resilient to automation trends; poorer regions have a larger number of jobs with greater technical potential for automation. In other words, it is likely that automation will intensify the UK’s already stark levels of regional inequalities.

Even as increasingly capable and inexpensive machines put downward pressure on middle and low-income jobs, technological change is likely to increase the income and quality of work of highly skilled labour in roles which augment machines. As work is reshaped by automation, we therefore risk becoming an economy of DeepMind and Deliveroo, one where technologies support awe-inspiring advances in our collective intelligence while also enabling an intense and exploitative form of ‘digital Taylorism’ to thrive.

The future is not technologically determined

Taken together, these trends – a rising share of national income flowing to capital, and a polarising labour market ¬– mean automation is likely to unleash a powerful dynamic of divergence. That is, if public action is not taken.

Managed well, automation could create an economy where prosperity is underpinned by justice. If the productivity benefits are realised and fairly shared, the growing use of AI and machines in the economy should enable us all to work better but less, in a more sustainable and materially abundant society.

Crucially, the future is not technologically determined; it will depend on the choices we make. If we want a society of shared plenty, we will need to build the economic institutions that enable us all to share in the gains of technological change. Our report sets out the steps that can help achieve this.

Using automation to raise productivity

First, we should seek a managed acceleration of automation. Given the UK’s poor productivity performance, in too many workplaces it is the absence of robots that is the problem, not their imminent increase. While the top one per cent of firms have seen average productivity growth of around six per cent per year since 2000, one-third of UK companies have seen no rise in productivity at all, in part due to slow rates of technological adoption.

Reaping the potential productivity gains of AI and robotics will consequently require a reorientation of industrial strategy towards the ‘everyday economy’. This means sectors such as retail, hospitality, care, and logistics where many people work, but too often in low pay, low productivity workplaces. We need a focus on accelerating the effective adoption of new technologies in non-frontier firms.

To this end, the more rapid adoption of digital technologies, including automation, should become one of the national ‘missions’ of the Government’s industrial strategy. To support this, a new partnership body, Productivity UK, should be established with the goal of raising firm-level productivity, with a particular focus on the adoption of digital and other technologies throughout the economy, beyond frontier firms.

The ethical use of robotics

Second, we need to accelerate automation on our own terms. The integration of robots into society – and the potential for our eventual eclipse by non-human intelligence – inevitably raises profound moral and ethical questions. As far as possible, these questions should be answered collectively, as a society-wide endeavour. The alternative is a dark prospect: leaving our future to be defined by the digital giants of Silicon Valley, who would monopolise the development and define the use of technologies.

A better solution is for society to proactively develop the legal, ethical, professional and behavioural framework that would govern the development and use of AI and robotics. We therefore propose the establishment of an ‘Authority for the Ethical use of Robotics and Artificial Intelligence’ to provide guidelines and recommend regulatory frameworks for the use and governance of these technologies.

New models of ownership

Finally, to ensure automation enriches all of society, we need to broaden and democratise capital ownership. Expanding the distribution of capital and pluralising models of ownership would help democratise who has a claim on the economic dividends of automation, both in terms of control and benefit. The IPPR paper sets out a series of strategies to dramatically broaden ownership of capital, from a Citizens’ Wealth Fund to employee ownership trusts to new models of profit sharing. At their heart is a belief that new and pluralistic models of ownership will be needed to ensure automation creates a future of shared economic plenty.

Taken together, these are the first steps towards a new institutional settlement that can better ensure technological change works for the common good.

This blog originally appeared in Political Quarterly.

Mathew Lawrence is a senior research fellow at IPPR he tweets @DantonsHead.