Having spent over 30 years in UK policing, including decades investigating complex fraud cases and the resultant proceeds of crime investigations, I’ve witnessed first hand the immense challenges that arise from managing disclosure obligations under the Criminal Procedure and Investigations Act 1996 (CPIA).
The complexities inherent in serious economic crime cases means that investigators and prosecutors can face mountains of hard copy documents, and terabytes of digital data, that must be meticulously reviewed to ensure compliance with legal disclosure requirements.
Indeed, I would cite disclosure as being one of the main reasons that fraud has effectively disappeared from the UK policing radar. It is one thing to take a statement, gather exhibits and effect an arrest; it is quite another to have to manage the obligations of disclosure. To put this into context, disclosure is significantly more resource-heavy than the actual evidence gathering.
AI – A call to action
A recent government-backed review, conducted by Jonathan Fisher KC, has identified the gravity of this challenge. His recommendations rightly highlight the desperate need to modernize our approach to criminal disclosure, through the integration of advanced technology, specifically artificial intelligence (AI). See also my article Opinion: Will common sense prevail in the UK’s long overdue review of disclosure rules?
As someone who has repeatedly grappled with the CPIA disclosure requirements, I wholeheartedly support this call to action. And as someone who is still undertaking investigation work all these years later, I readily admit that my grasp of AI and what it can do somewhat escapes me (it is as far as I am concerned rocket science that has been melded with brain surgery). Whoever can harness AI to streamline disclosure will forever be a hero to law enforcement and investigators in the UK, and will likely be enshrined in a statue by the Crown Prosecution Service (CPS).
I would cite disclosure as being one of the main reasons that fraud has effectively disappeared from the UK policing radar.
In today’s digital era, the sheer volume of evidence collected during complex fraud investigations is unprecedented. A single investigation can easily encompass millions of individual documents, along with countless emails, text messages, financial records and vast amounts of multimedia data. Historically, this has required painstaking manual review – an inefficient, costly and error-prone process. Despite investigators’ best efforts, manual disclosure processes have become increasingly untenable, creating significant risks of procedural errors that can derail entire prosecutions.
The scale of the problem is obvious, but the gravity of the current situation may be lost on some. Disclosure failures have already contributed to high-profile cases collapsing, eroding public confidence in the justice system. Jonathan Fisher’s review emphasises the urgency of reform, and his comment that we must “fight fire with fire”, using the very digital technologies to help tame the data explosion, is inspiring. In effect, he is not demanding evolution, rather revolution.
It must be remembered that the benefits of integrating AI into disclosure processes extend to the defence teams, and thereby the defendant. Under the current system, defence lawyers – particularly those representing clients on legal aid – often face immense challenges navigating voluminous disclosures which may bury the critical exculpatory evidence amid mountains of irrelevant material. Targeted, accurate, and manageable disclosure would significantly level the playing field for all involved.
I also welcome Jonathan Fisher’s proposed creation of a specialized working group, bringing together investigators, prosecutors, defence lawyers and judges to evaluate and select appropriate AI tools. I, for one, would welcome the opportunity to assist wherever possible. The benefits to my generation will likely be limited given our remaining shelf life, but the positive knock-on effects for those following us may be hugely impactful.
Investment in training
It is, of course, essential that any shift towards increased use of AI is accompanied by an investment in training. Many police officers lack the type of insight and training required to undertake voluminous and often technical disclosure exercises. My training was almost non-existent: instead I learned like my peers, through trial, error and experience. This must change.
Another point raised in the report is the recommendation for early discussions between investigators and prosecutors about disclosure strategies. This was always something we addressed from the outset of a complex case with the CPS lawyer overseeing the case. This significantly reduced procedural complications down the line, avoiding last-minute disclosure ambushes and disruptions.
The suggestion to reject the “keys to the warehouse” approach – where defendants have unfettered access to all gathered evidence, regardless of relevance – may also be prudent, albeit there is a debate that if the suspect knows where the evidence is to prove their innocence, why not let them show us early?
The argument goes that if there is significant evidence to undermine the prosecution case, it is better to know from the outset before embarking on costly disclosure processes. Although the risk of evidence being tampered with or destroyed is minimal where forensically imaged digital data is concerned, the counter argument is that a serious fraudster may have already planted and woven the very nuggets that would undermine the prosecution case into their scam as they developed it.
Trust in AI
Ensuring that AI systems operate transparently, ethically, and within clear guidelines will be crucial for maintaining trust in their application within the criminal justice context. The successful integration of AI technologies promises not merely to ease the administrative burden but fundamentally to strengthen the pursuit of justice. Prosecutors will benefit from increased efficiency, defence lawyers from more precise disclosures, and ultimately, the judicial system from enhanced fairness and reliability.
Embracing AI in CPIA disclosure isn’t just advisable, it is absolutely necessary. Jonathan Fisher’s timely recommendations offer a practical, realistic roadmap toward a more robust and fair justice system.
Tony McClements is Head of Investigations at BVI asset recovery law firm MKS Law.
