The Government is considering a retreat from the aspiration of opening large quantities of public sector data for re-use free of charge, HM Treasury has suggested in a discussion paper.
Looking at the value of data across the public and private sectors it says that questions of privacy and ownership make the ideal of universally free data unrealistic. “Rather than rely on an open/closed distinction, data access should be seen as a spectrum, with different degrees of data openness,” it proposes.
According to the Treasury, the paper, The Economic Value of Data, is part of a government-wide effort to develop a strategic approach to data, including its use in AI algorithms. This includes such initiatives as creating the Centre for Data Ethics and Innovation.
The Treasury’s main concern is in obstacles to realising the value of data. But it also appears to part company with the policy driven by the Cabinet Office over the past eight years that the optimum way to achieve this is to release datasets for free re-use.
While acknowledging that in some sectors such as transport open data “has had a clear economic and social benefit”, it says this does not mean that open is appropriate or beneficial for all forms of data.
As well as objections on the ground of security or privacy, it notes that where data is already monetised, making it open would remove a source of income.
“This could harm business models which use resale to invest in better data gathering, and in the public sphere, could result in taxpayers replacing lost revenue,” the paper says. The argument suggests that the trading fund model in which agencies make money from their data is back in favour – at least at the Treasury’s end of Whitehall.
No revenue sharing
The paper also comes down firmly against proposals to allow individuals to share revenues from businesses earning money from data relating to them. This would require a fundamental change in the law and create “substantial technical and legal challenges in seeking to value individual contributions of data, not least because personal data often relates to more than one person”, it notes.
Meanwhile, extending ownership rights to personal data “would ignore the fact that the overall value of a dataset cannot be divided equally amongst its constituent parts.”
Comments on the paper can be sent to: [email protected]
Image from GOV.UK, Open Government Licence v3.0