Proper disclosure by the prosecution lies at the heart of any criminal trial. Without it, unused material which assists the defence may never see the light of day, increasing the prospects of a wrongful conviction.
Recently, however, the exponential increase in digital material has left the system creaking under the strain (one current Serious Fraud Office case comprises 48 million documents). This has resulted in important material being overlooked during the disclosure process, cases collapsing and millions of taxpayer pounds being wasted. In turn, this has led to more time being spent on disclosure, while the growing mountain of digital material continues to overwhelm authorities. This is the disclosure conundrum – everybody wants the exercise to be done properly and yet it threatens to overwhelm other equally important aspects of the process (for example, the right to trial within a reasonable time).
The government turned to Jonathan Fisher KC to conduct a root-and-branch review of the system and to see where improvements could be made. His final report was issued on 20 March. He has left no stone unturned in his quest for a solution. In considering the extent to which he has succeeded, we consider three specific aspects of the review:
i) the manner in which schedules of unused material are presently compiled;
ii) the proposed paring down of schedules pre-charge and/or where a guilty plea is likely; and
iii) Fisher’s rejection of the so-called ‘keys to the warehouse’ approach (i.e. handing all the unused material to the defence and leaving them to review it for themselves).
Preparing schedules of unused material
The preparation of these schedules plays a vital part in the disclosure process. In the absence of an accurate description of the contents of an undisclosed item, the defence cannot tell if it may assist them. Going through these items manually, however, and producing the descriptions is hugely time-consuming and does not guarantee the accuracy of the schedule in any event. As Fisher points out, a 2017 review found that 25% of unused schedules in significant Crown court cases were ‘wholly inadequate’. These problems subsist even for the most well-resourced prosecutors, such as the SFO. That body often employs paralegals and/or junior members of the bar to undertake the exercise, when those individuals have variable levels of understanding of the case and/or little interest in it. At least one recent SFO case has been derailed in this way.
Fisher therefore considers the alternatives, the first being schedules which give details only of the metadata of each document, rather than a written description of their contents. In other words, information would only be given about the author of the document, the date of its creation, the dates/times of transmission, the identity of recipients and so on. Schedules of that description could be created by conventional software in hours, rather than the months and years presently taken to create schedules containing manual written descriptions.
Fisher suggests that metadata schedules might, in some instances at least, contain as much, if not more, information than conventional written description schedules. That sounds ambitious to us. There is only so much that can be gleaned from the metadata and defendants cannot be expected to recall the actual contents of communications years after the event, even with a prompt as to who might have authored a particular document and when.
Fisher also suggests that hybrid schedules could be produced, with written descriptions being used for documents of potential substance and metadata schedules being used for material of peripheral relevance. That suggestion merits further consideration, although two obvious issues arise. The first is determining which are the substantive documents and which are of lesser relevance. The second is how much time will be saved if written descriptions still have to be applied to all documents of substance. In the end, perhaps not very much.
Another alternative considered by Fisher is the use of artificial intelligence to ‘read’ the unused material and produce a written description of each item. One assumes that such software, if not yet available, will be soon. However, as Fisher notes, it is doubtful that any current AI tool could pick up on nuances apparent to the human eye. For example, it is sometimes the tone of an email or the use of punctuation (for example, exclamation marks) which reveals the underlying meaning of a communication. If this is not picked up by the AI software and an anodyne description of the item given in the schedule, an important document may slip under the radar.
As the software improves, these issues will no doubt ease but they may not be eradicated completely.
Paring down schedules pre-charge and/or where guilty pleas are anticipated
As matters stand, pre-charge schedules of unused material must be provided to the Crown Prosecution Service in all cases likely to end up in the Crown court. Much of this work is ultimately wasted. Fisher’s research suggests that in 2022/23, about 532,000 police officer hours were used to build pre-charge files, including schedules of unused material, only for the CPS to decide to take no further action. Investigators would no doubt argue that time is similarly wasted when a defendant pleads guilty in an overwhelming case.
There is scope for efficiency savings here but it is important that the baby does not disappear with the bathwater. If officers are relieved of the obligation to prepare pre-charge schedules, there is a risk of less attention being paid to unused material generally. If that happens, the CPS risks charging weak cases on a misinformed basis or, worse still, important unused material going by the wayside.
In relation to anticipated guilty pleas, this is also a slippery slope. First, it involves making subjective decisions on how likely a defendant is to plead guilty and making disclosure decisions on that uncertain basis. Second, early disclosure of unused material to the defence is an important aspect of advising defendants, particularly in a case where issues arise as to whether the elements of the offence are legally made out or not.
Whatever happens, if the need for a schedule in certain cases is to be dispensed with, there must be a mechanism for ensuring that the unused material has still been reviewed and any items of particular significance provided to the defence.
Keys to the warehouse
Whether or not Fisher would recommend this fundamental change in approach was the most keenly anticipated aspect of his review.
For the following reasons (with which we agree) he has decided against this route:
i) the prosecution abrogating its responsibilities of disclosure goes against the grain and philosophy of our criminal justice system;
ii) while the bigger defence firms may have the resources to undertake major disclosure reviews, most smaller firms would be overwhelmed, both in terms of personnel and IT; and
iii) the cost to defendants (privately paying or via the public purse) of undertaking such reviews would be prohibitive.
Fisher simultaneously acknowledges, however, that there need not necessarily be a binary approach to the issue. For example, if the defence wishes to review in its entirety only one aspect of the unused material (‘the keys to the cabinet’ approach), this could be beneficial to both sides. Again, we agree.
He also postulates that in single-defendant cases it might be possible for the prosecution and the defence to ‘share’ material originally belonging to the defendant (for example, the contents of a mobile phone or laptop computer). In this instance, investigators would still review the material to ensure that any sensitive material was removed, and would be obliged also to highlight exculpatory material which was uncovered. They would be relieved, however, of the burden of searching for disclosable material, or scheduling the contents of the device. Fisher recognises the risk of this being the thin end of the wedge in terms of the burden of disclosure being transferred to the defence. To mitigate against this outcome, he suggests that smaller defence firms could seek to contract out of this arrangement, such that disclosure was dealt with by the prosecution in the normal way. We see the force of that suggestion but question how it would work in practice. It could also lead to complaints from convicted defendants that their lawyers did not do enough to represent them.
Conclusions
In some respects, Fisher’s task was a thankless one and he should be congratulated on producing such a comprehensive, professional piece of work. In addition to the matters outlined above, he puts forward a number of other ideas as to how the disclosure regime might be improved, for example easing the burden of making redactions and increased training for disclosure officers.
These are important measures but, as we suspect he would agree, they are sticking plasters on a gaping wound. The real issue is how to cope with the vast amount of digital data that pervades most cases and threatens to subsume others completely. There is no silver bullet and Fisher does not suggest otherwise. One can only hope, as he says, that while advances in technology have been the cause of the problem, they also prove to be its solution. Given the pace of AI progress, this may happen sooner rather than later. Until then, however, issues with disclosure will inevitably continue.
Christine Braamskamp is a partner at Jenner & Block. Christopher Coltart KC is head of business crime at 2 Hare Court
No comments yet