Your location:

Naming - and shaming?

    The Coalition government’s drive toward open data and transparency has been one of its most conspicuous successes. We have seen greater accountability (thanks, for example, to publication of public spending over £25,000 and detailed local government spending), and enhancement of citizens’ understanding of their own communities (see for example the crime maps on Ever more datasets appear on the aggregator site These data are available for anyone in the private sector with a bright idea to develop innovative services; as such, they are likely to be an important part of the push for growth.

    There are, of course, limits, as I argued in an independent report for the Cabinet Office earlier this year. When a data release threatens national security, or the privacy of citizens, then hard thinking needs to be done about the risks and potential benefits. One particular issue has arisen with respect to the Ministry of Justice’s court data released in the last few weeks, which includes sentencing data at the level of individual sentences from 322 magistrates and crown courts in England and Wales, excluding the defendants’ names, but including details such as age, ethnicity, offence type and sentence (aggregated to help anonymise them). Juveniles are excluded from the data. Traditionally, although data of this type are public in the sense of being read out in open court, they have not been made easily available online.

    This has been quite properly hailed as important progress for open justice, but has also prompted an important debate between advocates of transparency and other stakeholders, including my colleagues on the Ministry of Justice’s transparency sector panel, Will Perrin and Paul Clarke.

    Will set the ball rolling with a quote in the Guardian datablog:

    “The data published is anonymised, flying in the face of hundreds of years of tradition of open courts and public justice.

    “The MoJ need to have an open and public debate about the conflict between the central role in our society of open public courts where you can hear the name and details of offenders read out in public and crude misapplication of data protection.”

    Will’s involvement with a local community campaigning site based around King’s Cross gives him a sharp appreciation of the value to a community of knowing the full outcome of a police investigation, including the result in court. It gives people confidence in the system of justice, but the current cost of getting information about who has been found guilty or innocent of which crime is prohibitively high for poorly-funded or unfunded activists.

    Paul’s response is to point out that there is a big difference between openness in courts, and putting names on the Web where they are easily searchable and more or less permanent. This has implications for the rehabilitation of offenders (less serious offences become ‘spent’ after a crime-free period, in order to allow offenders to rebuild their relations with their communities), as also detailed in a blog by Loulouk.

    The MoJ data could be used to provide a service to identify anyone with a conviction (e.g. compiling and selling blacklists of offenders, along the lines of the rather sleazy, but not necessarily illegal, blacklists that have circulated in the construction industry) – not a happy thought, if we care about effective reintegration of offenders into society. Of course, such a blacklist can be compiled already, but a readily-available database would lower the costs and increase coverage, enhancing the business model. Paul’s point is that the change in scale and speed is one of those quantitative changes so large that it amounts to a qualitative change.

    The Disruptive Proactivity blog argued that one way around that might be to impose terms and conditions on use, or perhaps use Web technology to slow down the interaction, or to demand evidence of a genuine local interest (thereby preventing large-scale scraping of the data).

    Everyone agrees with Will that a debate is needed about the role and extent of publicity, anonymity and search in the digital age. The practical obscurity that paper records afford – information is available, but hard to get – no longer protects data subjects. The question is whether we can exploit the opportunities of transparency without undermining principles of justice, offenders’ rights and the interests of society in encouraging rehabilitation.

    I must confess I am torn. Let me add to the debate three additional points which pull in different directions.

    1. Equity. Paul rightly argues that the increase in scale is a game-changer. However, it will also have the effect of making the current situation fairer, in that a canonical record will be easily and publicly available.

    At the moment, the Web’s coverage is patchy – some offenders (e.g. the woman who put the cat in the bin) will be prominent on Google forever more, others will quietly disappear when the local paper stops archiving its old material, and others are unreported and therefore safely anonymous. Some information is behind paywalls, some is free.

    Furthermore, there is a big difference between a neutral, objective database and other sources of crime information (e.g. local papers). The database eschews sensationalism, whereas other sources may not. Arrests are more interesting than not guilty verdicts, convictions more exciting than successful appeals. The database does not pick and choose.

    So in the current situation, there is no telling whether, or how prominently, information about a particular offender will appear on the Web. And even if online sources of crime information are largely accurate, that does not guarantee that the reader gets the full story (e.g. after an appeal).

    It may be that the increase in equity from an objective release of complete and accurate data by the Ministry of Justice could at least partially offset some of the issues surrounding rehabilitation. At least everyone will know where they stand.

    2. Cost. Disruptive Proactivity’s idea is a nice one in principle, but there are severe practical problems with it. Any attempt to slow down the gathering of data without impeding its flow to legitimate users will involve erecting some sort of computer architecture to make users accountable, or to make them identify themselves in some way. This could be quite lightweight, but the result would be a cost to the system itself. The beauty of the current transparency programme is that it imposes relatively small costs on government and no costs on data users. A system that placed even small costs on users would have quite large cost repercussions for government. The idea is certainly worth exploring for certain classes of data, as I have noted elsewhere, but in today’s current financial climate this is not what administrators and policymakers will want to hear.

    In general, cost does need to be taken into account. If the MoJ were routinely to name offenders online, there would have to be checks to ensure, for example, that reporting restrictions were respected, or that juveniles or other vulnerable people were not named, or that victims were not inadvertently identified by identifying the offender (e.g. in domestic abuse cases). These procedures would not be cost-free.

    3. Practice. Regulations and computer architectures can only dictate behaviour so far. Much also depends on social norms about how we deal with abundant sensitive data. We can certainly anticipate abuses and work towards eliminating them (perhaps, for example, through anti-discrimination laws). But this will require legislation, red tape, awareness-raising and other measures, which take time. We need to evaluate what goes wrong, and what right. As the Disruptive Proactivity blog put it, “being right slowly is better than being wrong quickly.” To me, that suggests a conservative approach to this issue, with incremental steps rigorously evaluated.

    So on balance I would follow Paul and argue for caution, but the argument is very finely poised. It’s a vital debate, which should spread beyond the open data and the criminal justice fraternities.

    Transparency will change our understanding of government’s roles and responsibilities. Dilemmas will emerge increasingly frequently. These are arguments we will have to get used to.

    Kieron O’Hara is a senior research fellow in Electronics and Computer Science at the University of Southampton, and has a DPhil in philosophy from the University of Oxford. His latest book, 'Conservatism', was published in May 2011 by Reaktion Books.

    Kieron O’Hara is a senior research fellow in Electronics and Computer Science at the University of Southampton, and has a DPhil in philosophy from the University of Oxford. His research interests are:
    the politics and philosophy of technology, particularly the World Wide Web; transparency and open data; privacy; memory; and conservative philosophy.

    Centre for Policy Studies will not publish your email address or share it with anyone.

    Please note, for security reasons we read all comments before publishing.


    Sam Smith - About 3000 days ago

    The system should be requestor blind, so that it shouldn't matter who is making the request, but simply that bulk requests are infeasible. That is fundamentally different to them being "accountable".

    Purely in financial terms, in the context of the costs of keeping one person in prison, the costs you highlight here seem rather trivial as to be irrelevant and effectively zero if designed properly (and still be effectively irrelevant if not). While that is not outside scope of consideration for any current financial year, it is, in the medium term, effectively solved.


    View Comments (1)
    Comment on This

    Centre for Policy Studies will not publish your email address or share it with anyone.

    Please note, for security reasons we read all comments before publishing.

    Kieron O'Hara - About 2993 days ago

    As I said in my blog, Sam Smith’s ideas are nice ones, and I certainly don’t want to close down discussion of them. If they can be pursued, so much the better – let’s see what works in the relatively new area of transparency.

    However, I don’t take the two points in his comment. First, I wasn’t intending to imply that a heavyweight identification architecture was needed, in order to make data users accountable. Agreed, on the most lightweight of his suggestions, something to prevent scraping would help mitigate the problem. It’s a creative solution, and worth exploring.

    But even that would involve a change of architecture, and a redesign. There would be an increase in the overhead for government suppliers of data. The current model of putting stuff out on under the Open Government Licence has the advantages of cheapness, elegance and ease. I sense little appetite in the transparency programme for developing more complex interfaces in the short term.

    So my point was not that Sam’s suggestion involved too heavy an architecture, but that any change of architecture would involve some cost, which needs to be taken into account. Will Perrin’s ideas (put the info out in the clear) and Loulouk’s ideas (don’t put the info out at all) both fit into the current architecture. The middle course is an extra step whose cost (in manpower and admin resources as well as money) would have to be assessed.

    Second, it is of course true that the cost of publishing information is much less than the cost of keeping a prisoner, but that is obviously a non sequitur – it is the job of the prison service to do the latter, but not the former. Even if Sam’s assumption that the only costs involved would be initial costs was correct, that money would still have to be found in a cash-strapped service. Personally, I feel that the marginal costs would not be insubstantial either, as data would constantly have to be checked and monitored. Sam’s point that proper design would help is of course true, but information system design, sadly, is not the long suit of government.

    So to reiterate, these are interesting ideas, but the cost of architectural innovation does need to be taken into account.