Some reflections on the EC High Level Expert Group Report on the impact of digital transformation on labour markets
Professor Maria Savona
Science Policy Research Unit (SPRU), University of Sussex
Digital transformations are creating great opportunities but also challenges for modern labour markets. Supporting and steering transformations while reducing disruption is a pertinent policy challenge. While discussions must address the technological anxiety and declining working conditions associated with advances in AI and automation, they should also consider how digital transformations could reduce unemployment and underemployment, and increase prosperity and inclusion at the European level.
To explore possible solutions to these policy challenges, the European Commission in September 2018 convened an expert group on the impact of the digital transformation on EU labour markets. The group included experts from the public and private sector, alongside a few academics. The HLG convened over five monthly meetings that provided opportunities to cross-fertilise our own expertise. I was able to share the results of our ESRC SDAI project (TEMPIS) on the effects of technological innovation on employment and self-employment in the UK’s local labour markets and discuss our views on making innovation more inclusive. We were asked to think outside the box, providing ground-breaking policy recommendations based on empirical evidence – no mean feat! These have fed into a final report, released on 8 April 2019 and discussed in Brussels at the High-Level Conference on the Future of Work.
Our policy recommendations covered three grand areas:
a) a skilled workforce, enabling digital skills and lifelong learning, career counselling and labour market intermediaries that reduce skill mismatches;
b) new labour relations, that prevent risks to occupational safety and occupational health, particularly mental health, equalise the administrative treatment of standard and non-standard work arrangements, and reinvigorating social dialogue;
(c) a new social contract, that ensures neutral social protection, create a Digital Single Window and redistribute the value of digital ownership.
The nine recommendations are summarized on the expert group’s home page.
In this blog, I focus on the policy challenges around redistributing gains from digital ownership, giving my own perspective on the recommendations we made in our report.
What is the value of data?
So far, the collection and appropriation of data by companies has gone largely unquestioned. People willingly provide their personal data in exchange for using an online service, as in a barter. And firms also benefit from collecting their employees’ data, uncompensated.
But there is an increasing need to consider – and measure – the value of data. Companies’ capabilities to innovate and grow are now determined not only by their investments in R&D, training, engineering, design and so on, but also their ability to manipulate and learn from the data they are accumulating. Advancements in data development and analytics mean that such information are increasingly seen and measured as ‘knowledge-based capital’ or ‘intangible assets.’
For example, the algorithms behind Artificial Intelligence (AI) are fed by a constant stream of data generated as workers carry out their tasks. Yet through machine-learning, this AI will eventually develop the capacity to replace these workers. So not only are they contributing, unrewarded, to this process, but its very progress may ultimately cost them their job, as the firm strives to increase its productivity.
How can this value be distributed fairly?
Thanks to the EU General Data Protection Regulation (GDPR), people now have much greater control over the data they release to firms. But there is still no regulation for tracking this immense stock of data, nor an easy way to enforce such laws.
To address this issue, one has to ask:
“Who gives data their value, and who owns (or should own) this value?”
Is it the individuals who generate the information, or “data capitalists” (who invest in ways to capture, accumulate, analyse and use it)?
We could think of ‘Data as Capital’ – as an asset (albeit an intangible one), the owners of which could be taxed by fiscal authorities. Such a taxation scheme could be implemented by new fiscal institutions operating at a supranational level, given that traditional, national tax bases are being increasingly undermined.
Or, if we think of ‘Data as Labour,’ those who create and process them could be paid a wage premium. This approach could yield a range of advantages, from improving the quality and quantity of data to increasing the productivity of AI systems.
A third, potentially more inclusive option we identified is to treat ‘Data as Intellectual Property.’ The ownership of data belongs with those who generate it – either workers within a firm or consumers / users of a firm’s services – and is treated as their intellectual property. If these data are then used by the firm to increase their (intangible) capital stock, this ownership must be recognised and paid an Intellectual Property Right (IPR). Through this different type of contract (on the basis of IPR rather than labour contract), firms would pay a license to use such data, but that license should be tax-free for the workers / consumers.
This third approach avoids the administrative burden of enforcing a digital tax, or shifting digital ownership. It also means that consumers who stop paying a firm for goods or services don’t relinquish their rights to the data they provided at the beginning and throughout. And a similar protection would apply to workers, on leaving a firm.
But it requires first and foremost a change in the way we think about the nature and value of big data, in a context in which data is considered the ‘new oil.’ A crucial distinction must be made: while oil is essentially a source of ‘rent’ for those who stock it, the ownership and subsequently the value of data is much more diffused. It must not only be tracked, but be recognised at time of provision.
In light of this, I recommend that the next generation of data regulations should include, for example, a General Data Tracking Regulation (GDTR) and a General Data Using Licence Regulation (GDULR).