Earlier this year, Camilla de Silva, the Joint Head of Bribery and Corruption of the Serious Fraud Office (SFO) gave a speech on:

• the future of corporate criminal liability;
• the SFO’s considerations when entering into a DPA;
• use of AI to assist with disclosure; and
• how the SFO will deal with privileged material going forward.

Corporate criminal liability

Ms de Silva highlighted that the current law on corporate criminal liability, which relies primarily on the identification principle, is an “inadequate model” for attributing criminal liability to a corporate. Under the identification principle the prosecutor must prove that the “directing mind and will” of the company (i) committed the acts amounting to the commission of the offence, and (ii) had the criminal intent to commit those acts. As Ms de Silva pointed out, the identification principle can be problematic as it discourages record keeping, incentivises the company’s Board of Directors to distance themselves from the company’s operations and is much more effective in establishing liability to small or medium sized companies than to large complex multinationals.

As an alternative, Ms de Silva proposed “the extension of economic criminal offences for corporations by a s.7 type ‘failure to prevent’ offence”.

Deferred prosecution agreements (DPAs)

Ms de Silva stated that under the DPA regime in appropriate cases the SFO will seek assurance that the company in question has genuinely reviewed its internal controls, policies and procedures regarding compliance and as necessary, adopted new or modified existing controls, policies and procedures in order to ensure it complies with all applicable anti-corruption laws and “most importantly” that these are actually embedded into the business. Ms de Silva stressed that the ultimate responsibility for identifying, assessing and addressing risks remains with the Board of Directors and is “a critical factor in any DPA discussion”.

When deciding whether to impose a monitor, Ms de Silva noted that the SFO will be informed by “the extent to which the programme of corporate governance enhancements is complete at the time of the DPA resolution and whether an ongoing independent review and sign off on-completion is considered necessary in the context of the deficiencies identified in the corporate compliance programme and corporate governance.”

In short, the SFO is unlikely to impose a monitor on a company which can demonstrate an effective compliance programme and robust governance systems.

Artificial intelligence (AI)

Ms de Silva also explained how the SFO had successfully used AI during the Rolls-Royce DPA to reduce the number of potentially legally privileged documents to be reviewed by independent counsel by 80%. As well as saving resource and tax payers’ money, Ms de Silva states that this also “resulted in a more accurate and consistent review”.

Legal professional privilege

Ms de Silva clarified that the SFO expects companies to enter into a dialogue about the basis and scope of any claim to legal professional privilege, the shape of its internal investigation and the timing of interviews. Although Ms de Silva stressed that the SFO is not interested in any material that is genuinely privileged, it does “reserve the right to question, probe and where necessary challenge” assertions of privilege which are in its view excessive as it did in its application for declaratory relief in ENRC.