A report from think tank Reform has recommended that the government should create a “seal of approval… which indicates that data quality is satisfactory and that biases within data sets have been accounted for”.
This data kitemark should be developed and administered by the Department for Digital, Culture, Media and Sport, Reform said. It should be similar to the O’Neil Risk Consulting & Algorithmic Auditing model – a system developed by mathematician Cathy O’Neil to assess potential bias in algorithms and artificial-intelligence programs.
DCMS should also develop a “data quality assurance toolkit” with which to test the data held by public-sector entities, the report recommends. Reform also advises that organisations across the public sector “should offer synthetic data sets, which they can share with others so that requests for data adhere to the right data standards in each organisation”.
The above are the first three of thirteen recommendations made by the report, which is called Sharing the benefits – How to use data effectively in the public sector.
The other 10 recommendations are:
- Technology manufacturers working in the government space should ensure that their products work with the relevant application programming interfaces used by the public sector.
- All systems bought by public-sector organisations should use open standards.
- Government departments should support data-led projects and help engage citizens in how their data is used in service delivery.
- Government departments should implement audit trails showing how personal data is used.
- The Information Commissioner’s Office should work with central government to develop programmes to train public servants on issues related to the handling of personal data.
- The ICO should work with expert bodies to provide resources that help public-sector entities understand data-protection legislation and their responsibilities.
- The government’s newly established data advisory board should work to understand and, ultimately, remove the barriers that prevent data being shared between departments.
- Once appointed, the new government chief data officer should assume responsibility for data-sharing policy.
- The Cabinet Office – and not DCMS – should take the policy lead on how individuals’ personal data is shared across government.
- Local authorities should be involved in creating data standards and systems.
The report’s conclusion said that “there is huge potential to analyse and share data to improve effectiveness and citizen interaction with public services”.
“The current cost of not sharing data is high,” the report added. “It can lead to gaps in service provision, repetition of information – which can sometimes be traumatic for service users – and poor outcomes. Data sharing is crucial to improve service delivery and outcomes, and its benefits can be espoused by all.
“However, the public sector has not yet managed to maximise these benefits. Currently, there are uncoordinated attempts to share data that do not have the appropriate data infrastructure to allow public-sector organisations to access the data that they need to deliver the best outcomes. For the government to live up to its ambition of creating joined-up public services centred around people’s needs, it will need to provide standards and protocols and focus on the construction of a stronger infrastructure around data.”
The full report is available here.