Artificial Intelligence in Investigations: meeting the challenges of 'big data' within the SFO and sharing ACi's practical in-house experiences

On Monday 20 May, RPC co-hosted with the Association of Corporate Investigators (ACi) a round-table relating to "Artificial Intelligence in Investigations". 

We were delighted to be joined by Richard Day, Head of eDiscovery at the Serious Fraud Office (SFO) and Trevor Dickey, Technology Specialist from GSK who shared their market-leading experience. Participants, including a number of leading in-house investigators across EMEA, then discussed the internal challenges of acquiring and embedding AI, core AI functionality in data review technology and AI related pitfalls and opportunities. 

The roundtable was moderated by Sam Tate, Head of White-Collar Crime and Compliance at RPC, with the input of Neil Fletcher, Head of EMEA at the ACi. 

A wide range of issues were discussed and three key points of interest follow:

• AI is an important part of the solution to investigators' "big data" problem. Investigations produce an ever increasing documentary burden and SFO investigations can now involve many tens of million documents, more than twice the number in even large recent matters, and in some cases more than the number of documents contained in the Panama Papers disclosure. As well as managing and sorting increasing amounts of data, AI now opens up possibilities for more quickly arriving at key investigation documents for triage purposes and reducing legal review time. For example, reviewing documents for legal professional privilege has traditionally been barrister intensive activity. AI can now be used to facilitate at least initial triage of privilege documents, which in the future is likely to support the greater efficiency in investigations sought by the SFO while also reducing the possible inconsistencies that human involvement always creates.
• ACi members' approach to obtaining budget and approval for AI based technology often benefits from engaging with other internal corporate stakeholders. AI has a wide-range of non-investigative uses across commercial organisations, from contract review, to compliance related management information, procurement modelling and financial crime investigations. Therefore, when considering adopting AI, many ACi member who participated in the discussions suggested that it is helpful to engage a range of in-house stakeholders and to consider the relevant technology as part of an overarching digital strategy.
• AI does not operate in a vacuum, it is part of an investigation workflow and as the expectations of AI increase, AI based tools benefit from a development plan. In particular, mapping the workflow which any AI based tasks will facilitate is as material to the technology's success as its functionality and the configuration between tools. In addition, given the changing needs of investigations, ACi members noted that it can be helpful to see any external provider's plan for the development and scaleability of their AI based tools before an in-house team begins to rely on the relevant tool's functionality. 
• If you would like further information in relation to any of the above, please contact RPC's White-Collar Crime and Compliance team at toby.lamarque@rpc.co.uk or Kate Winter at ACi, admin@my-aci.com