17 Aug 2020 Data as public good, and the public sector
By Parminder Jeet Singh
Understanding the nature of the emerging digital economy and society, and recognising people’s collective rights to their data, can help to determine how society’s data and digital intelligence-based roles should be divided across public, community and private sectors.
The currently dominant models of digital technology, economy, and society were born and developed at a place and time of ascendant neoliberal ideology, namely in the US of the decades of 1990s and 2000s. These models are consequently almost entirely ruled by the private sector, with practically no role for the public sector. Given the need for rapid innovation and disruption at the early stages of digital technology application, private sector leadership may have had some justification. But with the digital society structures becoming entrenched now, and increasingly dominating all sectors, public sector’s appropriate role in a digital society warrants assessment.
Key digital transportation data being largely in the hands of a few digital corporations, some cities in the US have considered handing over practically the entire public transportation sector to private management. Massive AI-based private education projects may push into oblivion the school system as we know it, and along with it also the educational authorities. Corporations holding health data are set to reorganise the health sector supplanting the role of public health systems. Digital corporations are developing smart city projects in which their control over city data converts into de facto governance of the city.
Not just provision of services, the very acts of public policy-making and governance would soon be impossible without access to society’s digital data. Most of such data currently remains a private resource of digital corporations. They may pro bono share some of their data for purposes of public interest, for example, Facebook’s »Data for Good« and Uber’s »Uber Movement« initiatives. But such sharing obviously happens on the whims and terms of these corporations, and follows their own interests. It can hardly serve as the basis for how public policy and governance are to be undertaken in the digital age.
Let us consider a city that is planning smart traffic management, which will require access to real-time commuting data that mostly is only available with Google. Would the city authorities have to beg Google for this data or, as the dominant data economy model becomes mainstream and »accepted«, have to buy it? Even more likely, they may have to let Google, or some such digital corporation, manage city traffic services. This will involve monopoly service fees and lock-ins. Leveraging their new position, as the corporation involved gathers ever more city data, it will use it to forever keep improving its services and increasing the fees. Such a situation of irretrievable lock-in and ever-deepening dependence on a private provider for a public service may prima facie look entirely untenable, but that is where we seem to be imminently headed. This traffic management example can be extrapolated to every single area of public sector work, from city planning, community development and welfare services to utilities management, education, health, agriculture support, and more.
A central role of community data for a whole range of ser vices that have traditionally been provided by the public sector points to the immense, and indispensable, public value of such data.1 It makes a compelling case for community ownership of this data. Such ownership can enable free access to community data held by private companies whenever needed for purposes of public interest.2 This arrangement in fact appears absolutely necessary, unless the public sector is soon to more or less collapse completely. While data required for directly providing public services can be called as a core public interest need, other kinds of public interests are also relevant. Two such further purposes of public interest that require mandated sharing of data are (1) to ensure an open and competitive market for digitally intelligent products and services, and (2) to support domestic digital industrialisation.3
Is the public sector ready for new data-based roles? An appropriate theory about such roles for the public sector, and the enabling policies and laws like on community data ownership, are certainly needed first. But equally important are the practical details.
Much of the change and restructuring will take place within existing public sector bodies and institutions, like those providing services of transportation, health, education, welfare, etc. These bodies will have to become adept at collecting and curating the required data from their existing activities, as well as privately held data that they will get access to under community data ownership rules. Competencies will have to be developed to convert data into necessary digital intelligence, and use it to provide intelligent public services (of course with the help of data scientists). Considerable skill development and upgrading may be required for public sector workers, including bringing in new technical skills. But, at its core, digitalisation and datafication of the public sector is not so much of a technical challenge as often feared as it is of strategic visioning and able management. Public sector workers should be able to adapt to new data-intensive work processes as successfully they did to computerisation in the public sector many years ago.
Some of the required public sector restructuring may be relatively intensive, even if undertaken gradually to accommodate human and other kinds of costs. Some public-sector roles may indeed become less important in the digital society, but many entirely new ones will emerge.
With industrialisation, the public sector acquired the important role of providing key industrial infrastructure. It should be taking up a similar role with regard to digital infrastructure. If any such new role for the public sector is hardly ever discussed it owes largely to the digital society’s birth and upbringing in a neoliberal environment. Global, vertically integrated digital corporations, spanning several sectors of the economy, internalise what are appropriately infrastructural and public-sector roles. Not only are the new digital infrastructure roles private right from birth – a creeping acquisition of existing public infrastructure roles is also taking place. An illustration of this is private digital currency initiatives like Facebook’s Libra seeking to take over the government role of managing currency as the token of value in economic exchanges.4
New digital infrastructure areas range from digital connectivity and basic computing facilities to cloud computing and data provisioning.5 As the very basis for intelligent production – of intelligent products and services – data is required for all important digital economy activities. Being in the nature of information, data is prima facie a non-rival good. Also, as data is combined with other data its value increases dramatically. This makes a case for provision of important data as a common infrastructure to all digital economy actors in any sector. The current digital economy model, however, is based on exclusive appropriation of society’s data by a few monopolistic digital corporations. They thus increasingly control the value chains in all sectors. Such exclusive use of the common resource of society’s data is the main reason for increasing concentration of digital power, and to a good extent also of increasing economic and social inequalities. Data sharing, or providing data as a common infrastructure, maximises the benefits that a society can derive from data. Sufficient open availability of key data is also the sine qua non for a competitive digital economy, and for reversing the damage being caused by concentration of digital power in a few hands.
The concept of data infrastructure is drawing increasing attention.6 This differs from the earlier open data movement, which consisted mostly of putting public data out in the open for anyone to use. Key data in different sectors mostly used to be with the public authorities; but today private digital platforms are the biggest holders of such data. Furthermore, digital society’s granular and intrusive digital data is of a nature that requires considerable protection against misuse. Such data has to be shared in a regulated and managed manner.7 Data infrastructures are designed for safe sharing of sector-wide data taken from different sources.
Command over AI is the new basis of economic power.8 Various national AI strategies rightly focus on data availability, which requires data sharing.9 They promote institutions like data infrastructures, data trusts, data exchanges and data markets, in order to ensure increased access to data for digital economic actors. Although mandated data sharing does get mentioned in some places, these national strategies mostly discuss voluntary data sharing. It is not explained, however, why the biggest collectors of data – digital platform companies – will on their own share or even sell their data when they consider maintaining exclusive access to data to be their main business advantage. In pussy-footing the obvious need for mandated data sharing, the frames of these national AI strategies seem to be tactically avoiding too direct a confrontation with the dominant political economy of the digital society, backed as it is by the most powerful global economic and political interests. But since effective data access and data sharing lie at the heart of any possibilities for AI and digital industrialisation, this weakness ensures that these Al strategies are doomed to failure in their current forms.10
Data infrastructures are not ordinary optional projects that can provide certain benefits; they constitute the very foundation of a strong domestic digital and AI industry, and ensure its openness and fairness. Data’s privatisation and monopolistic appropriation, on the other hand, is at the core of the dominant digital economy model. There is no escaping this paradox; it needs to be squarely addressed and urgently resolved.
Public data infrastructures have to be a key part of the new digital institutional ecologies. Most of them will be directly run by the public sector as a part of existing public departments or agencies in different areas, or will be operated by setting up new cross-sectoral agencies. Some data infrastructures could be managed in partnerships with non-profits or businesses, and others run privately as regulated utilities. Effective regulation for data markets is also required. Public sector capacities need to evolve for all these roles.
Public data infrastructures in different sectors – commerce, transportation, finance, tourism, agriculture, health, education, the labour market, and so on, are necessary to (1) deliver respective intelligent public services, and (2) robust private sector development, supporting a host of competitive digital businesses in each area.11 Data infrastructures play a central role in digital industrialisation, especially by nurturing domestic businesses.12 When intelligent products and services are competitively available, and lock-ins made difficult with effective data-portability laws, it enables better distribution of digital power across an economy and society, as well as globally. This can ensure the best value for consumers, and greater bargaining power for workers and other small actors in digital supply chains.
India is developing public data infrastructures in many sectors, ranging from commerce and finance to health, education and agriculture.13 The EU is creating data exchanges in the areas of transport,14 logistics,15 and health,16 and a common database of health images to support AI applications in healthcare.17 Similar initiatives are cropping up all over the world. Public data infrastructures will in time further specialise and evolve to provide not just raw or semi-structured data, but also its higher derivative forms. These could range from structured data and trained AI models to actual AI as a (public) service.18
Much discussion gets devoted nowadays to speculation over »AI versus humans«. But the most important political and economic question currently is: Who owns and controls society’s AI systems, or »systemic intelligence about us«? This is granular, real-time intelligence about – and thus nearly absolute power over – every niche and element of our socio-economic organisation. Is it with a handful of actors? Should we all not own it collectively? (Although uses of such digital intelligence, in many acceptable areas, will certainly need to be licensed under regulated conditions to private businesses for greatest productivity.) Our collective ownership over systemic digital intelligence about ourselves, as well as over the data from which it is derived, implies that a society’s data and digital intelligence are to be public goods.
Striking at the heart of the default dominant model, such a public goods perspective provides us with a new point of departure towards a digital economy and society that is just, fair and equitable. It will take forward the mixed economy and welfare state models that characterised the dominant post-war consensus,19 but have been upstaged by the neoliberal assault.20 The latter has employed the cover of rapid digital flux to gain much new territory with respect to society’s systems and institutions. If properly conceptualised, strategised and politicised, the same digital shift can in fact be leveraged to rehabilitate the pre-neoliberal consensus. This is because the key resources of the digital economy data and digital intelligence have some inherent features of a »social commons«.21 ν
Excerpted from Economic Rights in a Data-Based Society: Collective Data Ownership, Workers’ Rights, and the Role of the Public Sector, a paper published by the Friedrich-Ebert-Stiftung (2020) examining the fundamental changes taking place to economic structures, centred on the key digital economy resource of data, and exploring the implications of these changes for the public sector and its workers.
Notes
1. Public value refers to the value created by government through services, laws, regulation and other actions. https://www.themandarin.com.au/104843-measuring-public-value/
2. India’s Al strategy refers to mandatory data sharing for purposes of public interest , and some EU policy documents are also beginning to veer towards this view.
3. India’s earlier referred draft e-commerce policy proposes such data.
4. https://www.cnet.com/news/facebooks-libra-cryptocurrency-could-be-banned-in-india/
5. The EU has infrastructural projects in areas of high-performance computing and low-power micro-processors required for large data and Al applications. https://ec.europa.eu/11ewsroom/dae/document
6. https://www.stateofopendata.od4d.11et/chapters/issues/datainfrastructure.html
7.»Open data« is in general useful, with little potential for harm. Digital economy data provides granular intelligence on specific individuals and groups and may carry great potential for harm. It therefore cannot just be made open to anyone and everyone without protections.
8. Russian President Vladimir Putin has observed that whoever takes the lead in Al will become the ruler of the world. This corresponds to leadership in industrialisation in an earlier era. https://www.theverge.com/2017/9/4/16251226/russia-ai-putin-rule-theworld
9. UK’s Al strategy at https://www.gov.uk/government/publications/artificial-intelligence-sector-deal/ai-sector-deal; India’s at https://www. niti.gov.in/writereaddata/files/document _publication/NationalStrategy-for-AI-Discussion-Paper.pdf?utm_source=hrintelligencer; and France’s.
10. The paths adopted by the US, as the first starter, and China, which thoroughly fire-walled its nascent digital economy, are generally not available at this stage to other countries for digital industrialisation.
11. See the chapter on “Public data infrastructures” in the paper “Digital industrialisation in developing countries”. https://itforchange.net/sites/default/files/1468/digital_industrialisation_in_developing_countries.pdf
12. Ibid.
13. Ibid.
15. https://www.iru.org/resources/newsroom/eu-digital-logisticsplatform-puts-e-cmr-test
16. https://ec.europa.eu/digital-single-market/en/exchange-electronic-
17. https://ec.europa.eu/commission/presscorner/detail/fi/memo_18_6690
18. Al as a service« is an emerging business model. The public sector will need to move away from just using Al applications which co promises its hold over the value of very important data that passes through its hands and also specialise in providing some public-infrastructural Al services.
19. Somewhat arbitrarily, treating communism here as an exception.
20. Caught on the wrong foot by a bipolar digital world dominated by the US and China, EU leaders are beginning to think aloud in favour of such a »middle path« political economy for the digital society. https://www.politico.eu/article/germanyfalling-behind-china-ontech-innovation-artificial-intelligence-angela-merkel-knows-it/
21. http://datagovernance.org/report/data-a11d-data-intelligence-
Parminder Jeet Singh is Executive Director of IT for Change. His engagements have been in the area of ICTs for development, Internet governance, e-governance, and digital economy.
Photo: kentoh/Shutterstock
Sorry, the comment form is closed at this time.