Written evidence from JUST: Access
Introduction
- JUST: Access is a social enterprise on a mission to make justice more accessible through technology. Founded in 2016, when CEO Sophie Walker, working as a solicitor specialising in criminal appeals, struggled to obtain court transcripts of her client’s cases in order to advice properly on appeal.
- With the support of the Legal Education Foundation, Social Tech Trust and most recently Mercia Asset Management, JUST: Access has developed the only AI-enabled transcription tool specifically for the justice sector aimed at radically reducing the cost and improving accessibility of transcripts.
- These submissions will respond only the 3rd questions: What could be done to make information on court cases more transparent and accessible.
- While the principle of Open Justice places a presumption on court hearings to take place in public, in reality there are many barriers to attending a court hearing. From identifying when and where a case is being heard (the Courtserve website is designed for professionals and requires looking laboriously through pdfs to find the case number), to travelling to the court and finding the right courtroom. There is no way a member of the public could learn about a late change to the listing or an adjournment, risking a wasted journey.
- Even with the advent of online hearings, there are still difficulties gaining access by citizen journalists, students, journalists and so on, as raised by The Transparency Project and others.
- The easiest way of accessing information about a court case is by reading the judgment of the case. However, only a tiny proportion of cases that are heard by the court lead to a judgment being published.
- As set out below, our current processes of transcribing and publishing transcripts means that Open Justice remains an illusion, a laudable principle but one that means very little in practice.
- However, by better harnessing the technology already in use by the courts, it is possible to transform the principle of Open Justice, to make our courts more open, transparent, and fairer to all.
Brief background on court transcription
- In a Court of Record, where hearings are recorded, it is possible to order that an official transcript be created from the audio recording of the hearing. Judgments are either written by the judge or given ex tempore meaning an oral judgment is given by the judge at the end of the hearing. This is then transcribed by a court-appointed transcription company and published. In Senior Courts, i.e. High Court and Court of Appeal, transcripts of oral judgments should automatically be created. In lower courts, the parties can order a transcript if they want to appeal and are relying on a procedural, factual or legal irregularity.
- Court transcription can only be provided by around 6 companies who were awarded the contract by the HMCTS. For criminal courts, the country is split into geographic regions which each company is responsible for. For civil courts, the party can choose which transcription company they wish to use. The costs are in the region of £90-£200 per audio hour.
- Once a transcript is ordered, the audio recording is retrieved either from the court itself (where it is burned onto a CD-ROM and sent by post or courier to the transcription company), or via the court recording database known as DARTS (criminal courts) or DARS (civil courts). Following transcription, a draft copy is provided to the judge for their review, before a finalised or “sealed” copy is promulgated by the court.
Who loses out
- The Ministry of Justice: the government currently pays the high cost to transcribe judgments in the Senior Courts. Somewhat incongruously, the Judiciary also spend high sums each year on accessing these same judgments via legal publishing databases.
- Litigants: The cost of court transcripts are considerably is 20-65% higher than industry average. Further, litigants in criminal courts in the North East have to pay 50%more for transcripts than those in London, as the transcription company mandated for the London courts is considerably cheaper than the one mandated for use in the North East.
- All of us: While the publication of judgments is necessary component of our common law system, research in 2017 showed a dramatic reduction in the number of judgments being published each year (https://www.infolaw.co.uk/newsletter/2018/11/open-access-case-law-get/).
What can be done
- Transcription of court judgments remains an important mechanism to record the decisions of senior courts. However the high sticker price and the friction around their production and publication means their contribution to open justice can only ever be a marginal one.
- There are alternate ways in which justice can be made more open and transparent that cost only a fraction of the price.
- One method is to make the audio/video recordings of hearings more accessible.
- Even before Covid-19, important hearings in the High Court and Court of Appeal, and all hearings in the Supreme Court are recorded and streamed. These videos are made available for a period of time on their websites, and then are removed from the public domain.
- Now with more video or hybrid hearings taking place, the audio recordings of hearing are stored on a cloud-based server, it is even easier for such recordings to be used within the public domain, although access is heavily limited.
- A small dataset of UK Supreme Court videos from hearings in 2010-2011 is available on the National Archives website, and access is unrestricted (under the Open Government Licence). This demonstrates the “art of the possible”, that videos of court hearings can be released without re-use restrictions.
- Using our transcription tool customised for the justice sector, we are working on ways to make video and audio recordings of hearings more useful, such as creating smart bookmarks and keyword searches, so it is no longer necessary to watch the full hearing to identify the parts of the hearing that are most relevant to the issue you are interested in.
- Supplementing the published judgments with video and audio content of hearings would radically improve access to legal information for journalists, lawyers, legal researchers and members of the public.
- The rise of remote hearings provides an opportunity here the audio file of the hearing is much more accessible. While in the past, the audio recording of hearings was saved onto a CD-ROM and file sent by courier, now the video conferencing platforms such as CVP save a recording to secure cloud-storage system. Hearings that take place by Microsoft Teams also have the capability to recorded and an automatic transcript created.
- It is understood that there are concerns amongst Judges that automatic transcription is not accurate enough to be a substitute for the transcription prepared by the companies on the court transcription tender. The accuracy of automatic transcription can be greatly reduced where there is background noise, multiple speakers and complex vocabulary as can be expected within a courtroom. However, it does not follow that there is no place for automatic transcription in the courtroom, as the following use-cases demonstrate.
Use cases for automatic transcription
- A litigant-in-person at a Final Hearing in family court.
- In family courts the pressure on court time means that judgments are most often given ex tempore. Without a lawyer to explain the meaning of the judgment to the litigant, unrepresented parties can finish proceedings with little understanding of what decisions have been made about their children and more importantly, why. This may lead to unmeritorious appeals or be a further source of friction between the parties.
- Instead, the litigant could be provided with the audio recording of the judgment. Automatic transcription could be used to make the audio more accessible.
- A judge preparing a written judgment following a decision being appealed
- A judge required to prepare a written judgment of a decision if it is being appealed to an appellate court or tribunal. The creation of an automatic transcript from the audio recording of the hearing would allow the judge to identify the most significant parts of the evidence and listen back to these, rather than having to listen back to the entire hearing.
- Academic Researchers investigating fairness and effectiveness in outcomes of court cases
- While the Administrative Data Research UK and Ministry of Justice have created a dataset that contains basic information on the outcome of cases over a 5-year period, this is insufficient to undertake effective research to compare outcomes. Only by analysing sentencing hearings themselves would the researchers learn about mitigation and aggravation, or where a case sits on the sentencing guidelines. At £90-£200 an audio hour, transcription is cost prohibitive, but by accessing audio recordings and utilising automatic transcription, such insights can be unlocked much more easily and affordably.
- HMCTS making better use of its data
- As Dr Natalie Byrom has written of in her Digital Justice Report public in 2019 report and reiterated since, including to this committee about why better court data is vital to address the backlog created by Covid-19, basic management data is still not routinely collected on cases. For example, her evidence stated that HMCTS still “does not have robust data on the speed, cost and outcomes of different approaches to tackling the backlog.”
- Creating a data-rich data set using automatic transcripts of hearings would give HMCTS unprecedented insights into what is happening within courts and tribunals. It would provide an evidence basis for the Judiciary to understand, for example, whether Practice Directions they have been made are being applied correctly (Recommendation 11, Dr Byrom’s 2019 report), or the impact of pilots, interventions and other policy changes.
October 2021