Automation on our own terms

How AI and robotics can create a better world of work

By Benedict Dellot and Fabian Wallace-Stephens

Follow Benedict and Fabian on Twitter @BenedictDel @Fabian_ws

This article is an extract from the RSA report The Age of Automation: Artificial intelligence, robotics and the future of low-skilled work

The notion that AI and robotics will soon destroy large swathes of jobs does not stand up to scrutiny.

Science fiction vs. economic fact

We began this report by asking what advances in AI and robotics might mean for workers in the years ahead, particularly the 13.9 million people in low-skilled jobs. On the one hand is the alarmist viewpoint that says we are on the cusp of economic disorder. In the near future, some say, automation will be widespread and destructive, individuals and possibly whole communities will be displaced, inequality will accelerate to new heights, and well paid jobs will be few and far between — reserved only for the elites who own or manage the machines. Proponents of this argument point to the growing list of technological achievements, from software that can write news articles and machines that can shuttle goods around warehouses, through to algorithms that can manage logistical supply chains and ‘therapeutic’ robots used in mental healthcare.

No one is doubting the scale and pace of technological accomplishments, but the notion that AI and robotics will soon destroy large swathes of jobs does not stand up to scrutiny.

The UK unemployment rate is 4.4 percent, the lowest it has been since 1971 (see ONS for more details). There are now more people who want to work less hours than who want to work more (see below), and the redundancy rate is broadly on a downward trajectory.

For every article warning of technological unemployment, there is another complaining of skills shortages. Moreover, as our RSA/YouGov survey reveals, just 14 percent of the country’s businesses are currently deploying AI and/or robotics, or plan to in the near future. Contrary to colourful newspaper headlines, most employers are not paying attention to innovations in machine learning, deep learning or advanced robotics. The economic facts belie the sensational science fiction.

Does this warrant a collective sigh of relief? Only if we believe the status quo in our labour market is desirable — which it is not. As noted in the introductory chapter, and emphasised in Matthew Taylor’s Review of Modern Working Practices, the UK’s labour market performs poorly on a number of measures. While work is plentiful, the bulk of it is low-skilled and low-paid, and our productivity levels are abysmal. On average, UK workers are 35 percent less productive than their counterparts in Germany, 30 percent less than US workers and 9 percent less than Italian workers. Sluggish productivity growth has in turn been felt in stagnant wages, with real median wages still far below their pre-crisis levels. Workers are getting by on the pay packets of 2005.

Share of UK workers under and overemployed:

Source: RSA analysis of Labour Force Survey

Overall attitudes of business leaders towards technology:

Source: RSA/YouGov survey of 1,111 UK business leaders (Fieldwork conducted 10th-18th April 2017)

Our central argument is that the deployment of AI and robotics could help the UK forge a path towards a better world of work. These technologies could phase out mundane work, raise productivity levels, open up the door to higher wages, and allow workers to concentrate on more human-centric roles that are beyond the technical reach of machines. This is just as true for low-skilled workers as it is for high-skilled ones.

Most of the business leaders who took part in our survey agree with this sentiment, albeit when asked about technology in the round.

• Nearly half (43 percent) say that new technologies (including but not limited to AI and robotics) will lead to incremental automation and greater prosperity in the long run.
• Just 15 percent take a negative view that automation will be significant and that technology will harm livelihoods (see above chart).

Three risks

Yet there is no room to be complacent. As with all innovations, AI and robotics if deployed on a large scale would result in both losers and winners. Some geographic areas, demographic groups, occupations and sectors will be hit harder than others.

Drivers in the taxi industry will undoubtedly be rocked by the introduction of self-driving cars, just as builders and labourers would be disturbed by the arrival of robotics on construction sites.

While we are doubtful these machines will lead to the heavy loss of jobs, they are almost certain to result in their transformation. And while there are likely to be jobs created to replace those that are ultimately phased out by AI and robotics, people will be required to retrain, shift careers and move home in search of new opportunities.

There are three main risks of embracing AI and robotics unreservedly:

1. A rise in economic inequality — To the extent that technology deskills jobs, it will put downward pressure on earnings. If jobs are removed altogether as a result of automation, the result will be greater returns for those who make and deploy the technology, as well as the elite workers left behind in firms. The median OECD country has already seen a decrease in its labour share of income of about 5 percentage points since the early 1990s, with capital’s share swallowing the difference. Another risk here is market concentration. If large firms continue to adopt AI and robotics at a faster rate than small firms, they will gain enormous efficiency advantages and as a result could take excessive share of markets. Automation could lead to oligopolistic markets, where a handful of firms dominate at the expense of others.
2. A deepening of geographic disparities — Since the computer revolution of the 1980s, cities that specialise in cognitive work have gained a comparative advantage in job creation. In 2014, 5.5 percent of all UK workers operated in new job types that emerged after 1990, but the figure for workers in London was almost double that at 9.8 percent. The ability of cities to attract skilled workers, as well as the diverse nature of their economies, makes them better placed than rural areas to grasp the opportunities of AI and robotics. The most vulnerable locations will be those that are heavily reliant on a single automatable industry, such as parts of the North East that have a large stock of call centre jobs.
3. An entrenchment of demographic biases — If left untamed, automation could disadvantage some demographic groups. Recall our case study analysis of the retail sector, which suggested that AI and robotics might lead to fewer workers being required in bricks and mortar shops, but more workers being deployed in warehouse operative roles. Given women are more likely to make up the former and men the latter, automation in this case could exacerbate gender pay and job differences. It is also possible that the use of AI in recruitment (e.g. algorithms that screen CVs) could amplify workplace biases and block people from employment based on their age, ethnicity or gender.

There is every possibility that society will prevent AI and robotics from becoming mainstream because the dangers seem to outweigh the benefits.

When esteemed figures such as Elon Musk and Bill Gates warn of technological threats, the public and politicians undoubtedly listen. New technologies, including AI and robotics, will always create tensions and present new risks. But it would be a tragedy were we to lose sight of the enormous potential they also have for helping society address its biggest challenges — from managing an ageing population to lengthening lifespans to combating climate change.

Inclusive automation

The challenge, then, is to accelerate the adoption of AI and robotics but in a way that delivers inclusive automation — a kind that acts in the service of workers. How might this be achieved?

Many point to universal basic income (UBI) — an unconditional grant paid to every citizen — as the surest way to give people economic security in an age of automation. The RSA has itself been one of the strongest advocates of piloting UBI in the UK. Yet our response to technological disruption must encompass more than an overhaul of our welfare system. It is vital that policymakers, educators, regulators and others look across the lifecycle of technology, and intervene where necessary to encourage a positive outcome at each stage — from developing benevolent machines to equipping young people with modern skillsets. Effort must be made to:

• Develop benevolent machines — Programmers, tech companies and their investors should be steered towards developing benign forms of technology
• Accelerate the adoption of machines — Employers should be encouraged to deploy AI and robotics in a way that enriches rather than harms their workforce
• Future-proof the workforce — Educators must prepare future generations with the skills and competencies that will allow them to thrive in an automated economy
• Create a modern social contract — Our tax and welfare systems must evolve so that those who reap the most rewards from automation support those who lose the most
• Democratise the ownership of machines — The ownership of machines and the organisations that deploy them should be expanded so that more people can share in technological wealth

While it is beyond the remit of this article to spell out fine-tuned policy recommendations, here we take a tour through possible interventions under each of these headings (see below).

Some of these proposals are quick wins that can be implemented with little disruption or financial expense. Others are long-term, ambitious and will demand root and branch reform of our public institutions. There are also likely to be cases where an internationally coordinated response is required to manage the fallout of automation — not least when technology is being developed outside of our national borders.

In every circumstance, we should be guided by an overarching principle that it is never too early to begin planning for an economy where AI and robotics are ubiquitous.

Achieving inclusive automation through policy and practice:

Develop benevolent machines

The late economist Anthony Atkinson said that,

“too often technology is discussed as if it has come from another planet and has just arrived on Earth”

The reality is that society can and should shape the development of machines, including by eliminating potential flaws as they are being designed.

Progressive elements of the tech community should take a lead on drafting and signing up to ethical frameworks that would steer programmer behaviour, as the IEEE has done in the US. Philanthropic foundations and socially conscious investors also have a role to play by funding technologies that have more benign effects on workers.

More broadly, careers in the AI and robotics professions must be opened up to wider sections of society, so that technology is built with everyone’s interests in mind.

We advocate:

• Establish an ethical framework for AI and robotic engineers — Several large tech companies — including Apple, Amazon and Google — have committed to creating new standards to guide the development of AI (see Partnership on AI for more details). A recent EU Parliament investigation has followed suit in recommending the development of an advisory code for robotic engineers. These efforts should continue, but must not happen behind closed doors. Tech companies should use public engagement methods to canvass opinion on what society considers to be an acceptable and unacceptable use of AI, with a focus in this case on what happens within workplaces. Ethics training should be made a compulsory part of graduate computer science degrees, potentially culminating in a pledge akin to a Hippocratic Oath.
• Launch a national AI and robotics mission to boost the quality of work — The amount of funding flowing into the fields of AI and robotics is enormous. Yet much of this comes from private sources such as VC funds, and is often aimed at using AI for narrow commercial ends. The government should increase public spending on AI and robotics from its relatively low level (note our departure from the EU may cut off valuable streams of funding from supranational bodies). Part of this funding should be used to launch a new mission that rewards researchers developing machines that boost the quality of work, for example cobots that augment human labour. The mission could be organised through prize challenges, along the lines of the $4.5 million AI challenge just launched by the XPRIZE Foundation. The UK government should look to partner with likeminded countries on such an initiative.
• Mobilise the social investment community to sponsor benevolent AI and robotics — It is not just for the government to invest in socially responsible technology. Philanthropic foundations and non-profits also have a role to play. The Laura and John Arnold Foundation, an NGO based in the US, recently sponsored the development of a new algorithm to be used in criminal court proceedings — one that ignores factors like race, ethnicity and geography to ensure neutral assessments of defendants. In the same vein, non-profits in the UK should consider sponsoring AI and robotics that enrich the worker experience, such as recruitment algorithms that help employers find and hire workers from marginalised groups. A more significant step would be to establish a new social investment fund to back benevolent technology. Google has recently launched a new work initiative to fund tech-based innovations that will help people prepare for the changing nature of work.
• Open up pathways for marginalised groups to enter careers in AI and robotics — The tech community lacks diversity. Women make up just 17 percent of IT professionals and only 16 percent of new graduates from IT related courses, compared with 44 percent of new graduates as a whole. The Royal Society estimates that black and ethnic minority groups are over-represented in the ‘digital/IT sector’ but are underrepresented at senior levels. When machines are only built by a small group in society, they will ultimately only tackle the problems of a small group in society. Tech companies should redouble their efforts to recruit a more diverse cohort of programmers and managers, for example by partnering with groups like Code First Girls and InterTech LGBT.

Accelerate the adoption of machines

If the UK is to have a higher performing, higher paid labour market, then businesses and public services will need to ramp up their investment in AI and robotics. Particular attention should be paid to raising awareness of new technology among smaller organisations, most of whom lack the resources to investigate how AI and robotics might benefit them. Just 4 percent have embraced AI and/or robotics, compared with 28 percent of large firms.

Getting machines into organisations, however, is only half the battle. New technology must be integrated into organisational work practices and culture, which in turn requires workers and middle management to buy into the value of innovation. Accelerating the adoption of machines across the economy and among businesses of all shapes and sizes will demand new institutions, new incentives and new management practices.

We advocate:

• Establish a National Centre for AI and Robotics — The government should consider establishing a National Centre for AI and Robotics, or a Catapult centre of the same name. This would be tasked with increasing the diffusion of these technologies throughout the economy, for example by running trade shows where businesses and public services connect with technology firms; informing journalists of new developments to ensure more accurate reporting of AI and robotics; overseeing Knowledge Transfer Partnerships that place AI and robotics engineers within firms, particularly small ones; and canvassing the views of businesses so that researchers have a better understanding of their needs. A National Centre could also coordinate the aforementioned national mission to use AI and robotics for the advancement of good work.
• Encourage organisations to co-create automation strategies with their workforce — The LSE’s Mary Lacity and Leslie Willcocks find that technology is more likely to be integrated in organisations when the C-suite (the most senior executives) are actively involved in spelling out the benefits, engaging staff in how the technology should be used, and articulating the direct benefits to them. Wherever possible, businesses and public sector organisations should co-create automation strategies with their employees and the unions, and help workers retrain and pivot into new roles should machines take away some of their workload. Employers should also think carefully about which machines they purchase, as many can achieve the same outcome while having noticeably different effects on workers. Inspiration can be taken from Aviva’s decision earlier this year to consult its insurance staff on automation and retrain anyone who feels their job is under threat.
• Improve the financial incentives to purchase new technology — A small but meaningful proportion of businesses (14 percent) say they have not adopted AI and/or robotics because the technology is too expensive. There are several ways the government can make investment more affordable. First, as the CBI has suggested, the Annual Investment Allowance, which writes off 100% of qualifying capital expenditure (including tools, equipment and software) against taxable profits, should be doubled to £1 million. Second, Local Enterprise Partnerships should offer to knit together consortia of businesses to buy new machines in bulk with accompanying discounts. And third, the government should look at changing the rules on business rates, so that no plant and machinery investments are included in tax calculations.
• Rationalise and clarify data protection rules that impede tech adoption — The UK government, and where applicable the EU Parliament and Commission, should continue to review data protection regulation to ensure it does not unnecessarily discourage organisations from deploying AI and robotics. Current rules on data management appear ill suited to the use of machine learning algorithms that rely on large datasets. The EU General Data Protection Regulation coming into force in 2018 includes more robust requirements on organisations to gain consent from individuals to use their data. But as the tech company ASI Data Science points out, machine learning is concerned with finding multiple new uses for existing datasets, some of which may not be apparent when an individual is first asked for consent in the use of their data.

Future-proof the workforce

As machines become more sophisticated, so must the UK workforce raise its game. To thrive in an age of automation will require people to do one of three things: play a part in creating the technology, find a way of working alongside it, or identify a niche career that remains beyond the scope of it.

Educators need to encourage more students to enrol on computer science and STEM subjects, while simultaneously cultivating the deeper qualities of creativity, entrepreneurialism and overall grit that will help young people navigate their way through a turbulent labour market. In addition, lifelong learning must be given greater prominence and backing, such that adults have the wherewithal to transition between jobs and careers.

All of this will require a re-imagination of our educational institutions and the trialling of new training funds.

We advocate:

• Promote lifelong learning and pilot personal training accounts — Half of all workers in the lowest socio-economic group have received no training since finishing formal education. To help more people participate in lifelong learning, the government should consider piloting personal training accounts along the lines of those developed in France and Singapore. These would provide an annual credit of a few hundred pounds for workers to spend on any training course provided by accredited institutions. To pay for this, the government could reconfigure the Apprenticeship Levy into a wider ‘skills levy’, as suggested by the CIPD and the Taylor Review. A portion of the funds should be earmarked for marketing, to ensure demand for learning meets supply. In addition, the government should consider redirecting more funding to FE colleges, which are well placed to support lifelong learning efforts among low skilled groups.
• Experiment with new schooling methods that build soft skills resilient to automation — Advances in AI and robotics could lead to an expansion in human-centric occupations. A recent study from the US found that the share of the workforce in jobs requiring ‘high social skills’ grew by 10 percentage points between 1980 and 2012. To prepare young people for such roles, the government and educators should expand new schooling models that nurture the aptitudes of problem solving, critical thinking and entrepreneurial mindsets. High Tech High schools in the US revolve around project based learning, with each student required to undertake an internship within their community. Similarly, Studio Schools in the UK prioritise the development of emotional intelligence, communication skills and relationship building.
• Create sector roadmaps that anticipate and prepare for automation — The government and sector skills councils should form roadmaps for each major sector to understand the skills and jobs that are becoming more sought after, as well as those which are most at risk of automation. This should be informed through a live and continually updated index showing key developments in AI and robotics, as advocated by MIT’s Erik Brynjolfsson. The sector roadmaps would help schools, FE colleges, universities and other educators to better prepare their students for the offices and factories of the future. At a national level, the UK could follow in the footsteps of Germany’s government, which is considering the publication of regular reports on the changing world of work. Both the sector roadmaps and world of work investigations would require greater collaboration between employer groups, tech companies and educators.
• Modernise computer science courses and teaching methods — The demand for data scientists, programmers and system engineers is growing at pace. Yet 13 percent of computer science graduates are still unemployed six months after graduating, compared with 8 percent across all subjects. Evidence suggests this is due to a mismatch between what is taught in schools and universities and what is needed by the industry. The government’s new Digital Strategy goes some way towards addressing this problem, with a commitment to create generous new bursaries for computing science teachers that would improve the quality of courses. But there is also a need for more modular and fast-evolving training programmes outside of formal education, for example Makers Academy (see below). The government should offer the same subsidies to these fast-track courses as it does to HE computer science courses. Effort should also be made to link learners with jobs in the technology industry. In this regard, we welcome the government’s new Digital Skills Partnership, which will see tech businesses work with local government to help people move into digitally-focused jobs.

Makers Academy

Makers Academy describes its mission as ‘to teach as many people as possible how to create amazing products using beautiful code.’ It runs an intensive training course spanning a period of 3 months, with the intention that people can switch career into software development without returning to university.

The course planners update the curricula after every cohort, and aim to nurture software developers who are as good at collaborating and communicating as they building stable, fast and elegant products. The Academy has trained nearly 1,500 people so far, between 1/3 and ½ of whom are women. Graduates have found positions in Deloitte Digital, the Financial Times, HSBC and Thoughtworks, among many other organisations.

Create a modern social contract

To the extent that AI and robotics puts downward pressure on wages or eliminates jobs, it will push some workers into financial hardship. This demands a rethink of our social contract, broadly defined as the division of rights and responsibilities between workers, the state and employers.

In the medium to long-term, the government should consider the merits of adopting a Universal Basic Income (UBI) — a modest sum of money paid to every citizen on an unconditional basis. In the short-term, the government must ensure that the welfare system aids labour market flexibility while guaranteeing a minimum level of security for workers. Denmark’s ‘flexicurity’ system is one model to draw inspiration from. A consensus also needs to be built around tax reform. If machines do become more important as a source of income in our economy, it is reasonable to shift some of the tax burden away from labour and towards capital.

We advocate:

• Make ‘flexicurity’ a core tenet of a new social contract — Our social contract must evolve to meet the needs of a changing labour market. Against the backdrop of automation, our ambition should be to give workers a more robust safety net, while retaining the flexibility that encourages employers to take on workers. Much can be learned from Denmark’s ‘flexicurity’ model. Here, employers have greater freewill to hire and fire employees, but workers are entitled to up to 90 percent of their previous salary as they search for jobs, and are supported by a generous training regime co-designed with unions. The result is that a quarter of Danes in the private sector switch jobs every year, arguably leading to better job matching. Shifting to a model of flexicurity is ambitious: it will demand a restructuring of our institutions (with a bigger role for trade unions), and significantly more money to be spent on education and retraining. However, the potential prize may merit the size of the investment.
• Launch a meaningful pilot of Universal Basic Income — Universal Basic Income is often presented as a silver bullet that would help workers survive in the event of large scale job losses. We do not believe this scenario is likely, nor in any case that UBI would be the singular solution. However, UBI could be an important weapon in our armoury of policies to manage the modest labour market disruption we can expect from automation. It would allow workers to dive back into learning, give them more bargaining power vis-à-vis employers, and enable them to meet caring responsibilities (so responding to a demographic trend as well as a technological one). The government should put UBI to the test by facilitating a pilot in a UK town or city, along the lines of the experiment in Finland and Holland. The RSA welcomes the recent news that Scotland’s devolved government has committed funding to pay for basic income experiments among its local authorities.
• Move the tax burden away from labour and towards capital — How might these welfare reforms be paid for? Earlier this year, Bill Gates put forward the suggestion of a ‘robot tax’, which would charge businesses for deploying machines that displace workers. Yet it is difficult to see this working in practice, not least because it is impossible to distinguish between robots that complement humans and those that displace them. Nonetheless, the government should embrace the underlying principle of shifting the tax burden away from labour and towards capital. This would counteract any increase in the share of national income flowing to capital owners, which is caused as a result of automation. The government must keep the ‘employment wedge’ — the non-wage costs of taking someone on as an employee — as low as possible. Taxing wealth is notoriously difficult, given the opportunity for capital flight between countries. The OECD recommends recurrent taxes on immovable property as the least harmful to economic growth.
• Remove the financial obstacles to geographic mobility — The degree of technological disruption will vary from place to place. While some towns and cities will suffer declining job numbers and falling wages, others will see rising prosperity from the deployment of new machines. It is incumbent on the government to help people move closer to where new and better jobs arise. Particular support should be given to low paid groups who have fewer assets to finance a relocation. Building affordable housing in areas with plentiful jobs must be the priority, however there are other more immediate steps the government can take to aid geographic mobility. One of these is to reduce Stamp Duty, which can discourage or prevent people from purchasing homes in areas where jobs are in greater supply. The government should also undertake a feasibility study of relocation vouchers, a system of subsidising workers as they search for jobs in other towns and cities.

Democratise the ownership of machines

A final policy consideration is who owns the machines. Whereas the redistribution of wealth requires tough choices on tax changes and is subject to evasion, giving workers ownership over the technology that creates wealth is a cleaner solution that avoids connotations of dependency. A publicly owned sovereign wealth fund could be set up to invest in company assets and emerging technologies, and channel dividends to every citizen in the form of a ‘technological inheritance’. The fund could be built in the first instance by siphoning a percentage of capital stock from every IPO, possibly underpinned by short-term corporate tax relief so as not to discourage flotations in the UK.

Less radical but no less important, the government and business groups should take steps to expand the employee ownership and profit sharing movements, where workers have a direct stake in companies and by extension the machines they are investing in. This also means championing cooperatives where workers fully own and manage their organisations on a one-person, one-vote basis.

We advocate:

• Draft a blueprint for a UK sovereign wealth fund — Sovereign wealth funds (SWF) act as collective investment vehicles owned and managed by nation states. Today there are around 80 funds in existence, the majority of which were established after 2000. Through its investments, a UK SWF would give workers a stake in technology and a share in the companies that benefit from automation. One option advocated by a growing number of economists is to form a fund by siphoning a portion of shares listed in every company IPO (it would not have to be limited to technology companies). This could then pay out dividends to every citizen once it has reached maturity, whether in the form of a continuous dividend (as is the case with the Alaska SWF) or a one-off grant (what might be called a ‘technological inheritance’). The government in partnership with an alliance of civil society groups should begin drafting a blueprint for a UK SWF along these lines.
• Expand company profit sharing schemes — A more direct way to spread ownership of machines is by expanding company profit sharing initiatives. This would be of little use to those who lose their job to technology, but it would boost the incomes of the vast majority who remain in post. Only 8 percent of UK workplaces (with 10+ employees) are thought to operate profit sharing schemes. The government could raise this number by improving tax incentives, simplifying ownership frameworks, and establishing new rights. The Employee Ownership Association recommends streamlining employee ownership legal models from 5 to 2. The government could also establish a new ‘Right to Own’ rule — as the Labour Party has suggested — which would give employees of a company first refusal on purchasing shares up for sale.

· Champion cooperatives that turn workers into owners — Whereas employee ownership gives workers a stake in a company alongside directors and shareholders, cooperatives are wholly owned by workers. Each person has one share and one vote, and profits are typically divided equally among staff. Viewed in the context of automation, the advantage of the cooperative model is that workers can keep more of the wealth generated by adopting new machines. At present, just 2 percent of the UK’s GDP is accounted for by cooperative activity. Central and local government should look to turbocharge coop growth through practical interventions and financial assistance. This should include providing funding for coop incubators, as New York City Council has done through its new Worker Cooperative Business Development Initiative.

Two caveats on policy and practice

Above we have presented a number of policy and practice responses to an age of automation. This includes developing an ethical framework to guide the work of AI and robotics engineers; encouraging non-profits to invest in benevolent technology that enriches the worker experience; establishing a Centre for AI and Robotics that encourages greater take-up of innovations among industry; creating personal training accounts that aid lifelong learning and help workers as they jump from job to job; shifting the burden of taxation away from labour and towards capital, which is becoming an ever greater source of income; and drafting a blueprint for a UK sovereign wealth fund that would invest in emerging technologies, and give every citizen regular dividends or one-off grants in the form of a technological inheritance.

As worthy as some of these ideas are, however, they will never break through into mainstream policy discussions until they are seen to have legitimacy in the public eye.

Political history is dotted with U-turns on sensible interventions that did not receive public backing, such as Child Trust Funds, the proposed rise in National Insurance contributions for the self-employed, and more recently the move to reorganise social care payments. While our YouGov poll of business leaders reveals surprising support for some changes — such as 31 percent backing the idea of a UBI and 34 percent supporting greater employee ownership — many minds are clearly wedded to the status quo (see below).

Anyone championing a more inclusive automation must therefore step up advocacy efforts and begin forging new alliances.

A second major caveat when discussing policy and practice interventions is that low-skilled work will always be with us. Implicit in many of the discussions on how to manage automation is an assumption that if only workers can retrain and rise through the ranks, they will stay ahead of machines and see an improvement in their living standards. This is also the conceit that underpins the concept of social mobility. Yet it is of course impossible for all the UK’s 13.9 million low-skilled workers to move into higher skilled positions.

The task facing policymakers is therefore not to ‘save’ people from low-skilled work, but rather to make low-skilled work more financially secure and fulfilling in the long run. If education has its limitations, and tax and welfare changes prove too politically unpalatable, then the priority should be to distribute asset ownership more widely.

More broadly, there is a conversation to be had in society about what forms of work we value most, which is after all what influences how much consumers are willing to pay for different goods and services. While it is easy for economic pundits and think tanks to laud the move to a human-centric economy of caring, teaching and creative work, we must also be willing to pay for it.

Business leader support for different interventions as a means of managing technological disruption:

Source: RSA/YouGov survey of 1,111 UK business leaders (Fieldwork conducted 10th-18th April 2017)

We have argued that AI and robotics could be a blessing to workers rather than a curse. Implemented in the right way, new machines could raise productivity levels, phase out mundane work, boost flagging living standards, and open up the space for more purposeful and human-centric jobs to prevail. Equally, however, the onward march of technology could put downward pressure on wages, lead to greater monitoring in the workplace, and exacerbate economic, geographic and demographic inequalities.

The point is that technology is a tool to be wielded by society, rather than an independent force with a mind of its own. Whether or not AI and robotics helps or hinders workers will come down to the choices we make as employers, policymakers, consumers, investors and the wider public.

Yet this debate will continue to be a red herring unless we see a greater take-up of AI and robotics across our economy. The great irony at the heart of the frenzied speculation of whether new technology will lead to mass automation is that very few businesses are even embracing it.

Our poll finds that just 14 percent of business leaders in the UK are currently deploying AI and/or robotics, or plan to in the near future. Moreover, sales of industrial robots to the UK decreased between 2014 and 2015, with the UK purchasing fewer robots than France, the US, Germany, Spain and Italy. This mirrors a broader picture of chronic business and public sector underinvestment in capital expenditure, which goes some way to explaining the UK’s substandard productivity rates.

The RSA therefore calls for an acceleration in the take-up of this technology, but on terms that deliver an inclusive kind of automation which enriches rather than diminishes worker livelihoods. This cannot be taken for granted, and there are clear dangers of over zealously embracing technology with little regard for the consequences.

To ensure AI and robotics continues to work in our favour, we highlight possible interventions that can be made at every point in the technology lifecycle: funnelling more investment into socially beneficially technology; equipping the workers of the future with relevant skillsets; creating a more nimble welfare system that can accommodate greater labour market flux; and scaling up employee ownership models where workers have a stake in company profits — among other ideas.

The RSA and other organisations will continue exploring the impact of AI and robotics on the world of work — and rightly so. But this should not distract us from addressing the social, economic and environmental problems of the here and now. And there are many: a climate that is being irreversibly changed, public services under untold pressure, and the prospect of severe labour shortages as a result of our EU departure.

If anything, technology could be one of humanity’s most powerful weapons in resolving these issues — whether it is the use of robotics to relieve our strained social care system or the application of AI to identify new antibiotic treatments. Indeed, while the question on many people’s lips is whether we can live with these new machines, a more pertinent one to ask is whether we can live without them.

To find out more about our research, please contact Benedict Dellot

For full references and bibliography please visit the RSA website to download the full report