AI tools have reached prevalence—and this is of course also true for law enforcement and regulatory authorities. Regarding the use of such resources in internal investigations, a certain degree of reluctance has been seen at government authorities in the past—by no means at all of them—due to a lack of any legally established standards. However, practicioners engaged in internal investigations have long since outpaced lawmakers and developed their own best practices, with transparency and documentation being key factors in this regard.
Added value for the company and the authorities
Law enforcement and regulatory authorities expect from internal investigations a tangible benefit for their own inquiries. That is achieved by the company and its lawyers investigating the facts—normally more quickly and thoroughly than the authorities themselves would be able to. Yet the significance of internal investigations is not limited to fact-finding. At least just as important is the message they convey: the company demonstrating both internally and externally that it is not just paying lip service but has actually integrated compliance into its corporate culture. This is key when it comes to communicating with authorities—and, simultaneously, the yardstick against which the use of AI must be measured.
There are numerous reasons for using AI in internal investigations: it can automate a significant portion of manual reviews, prioritise assessments to be done by lawyers and reduce time spent drastically. Thanks to a deeper contextual understanding, relevant facts are identified that might have gone undetected if only search terms were used. At the same time, the rate of false-positive results decreases, which improves the efficiency of the entire investigation. This dual benefit—for the company, which saves resources, and for the authorities, which achieve better outcomes—must be comprehensibly and convincingly communicated.
Transparency and documentation as foundation
One of the first steps in every internal investigation is to define the subject and scope of the fact-finding process, as well as the specific methods to be used, in a 'terms of reference' framework. On that basis, the overall concept can be presented to law enforcement and regulatory authorities, as well as other stakeholders. This has always been the case and generally will not change in the future either. What is new, however, is that the use of AI and its significance for the internal investigation must also be documented. This does not mean that the inner technical workings of a large language model need to be explained down to the last detail—just as one does not need to understand how a processor works in order to rely on the functionality of a word processing programme. What is crucial, rather, is that the approach to the investigation is clearly documented, including the strengths and weaknesses of the methodology and, above all, its limitations.
In concrete terms, there are three levels to consider: first, the selection of the AI tool and the rationale behind it, including the security architecture and compliance with data protection requirements. Second, the concrete application—in particular the search parameters and prompts used, as well as the adaptation of the models to the specific facts of the case. Third, quality assurance, for example through spot checks, benchmarking and the involvement of technical experts. Law firms and companies that can provide seamless documentation of these three levels establish a solid basis of trust with authorities. This is naturally easier and more convincing with solutions developed in-house. Otherwise, even results obtained using sound methodology may be called into question.
From voluntary exercise to due obligation
Authorities' expectations have already shifted significantly with the growing prevalence of AI tools. What was regarded as an innovative pilot scheme just a few years ago is now increasingly seen as an indispensable part of every professional fact-finding process. Eventually, lawyers will no longer have to explain why they use AI—instead, they will have to justify why they are not. An internal investigation that fails to make intelligent and tailor-made use of AI risks falling behind best practices in the long run.
For lawyers, this means they need to have an even greater understanding of, and genuine interest in, AI and legal tech tools than before. This does not require a new type of lawyer. Rather, it gives prominence to the lawyer's original task of fact-finding—with the fundamental difference being that, besides the facts, the focus is increasingly on the methods used to deal with them. The ability not only to justify the use of AI to the authorities but also to effectively communicate it as a hallmark of the quality of an investigation is becoming a core competence in internal investigations. The lawyer who embraces this change early on not only secures a methodological advantage but also increases the authorities' confidence in the results of his or her own work.
