Automation, accuracy, risk and liability

HSBC's £64m fine by the FCA for weaknesses in its automated money laundering detection systems highlights the risks around automation that businesses and their providers need to be aware of.

Digital transformation strategies often involve the automation of existing processes, mapping, optimising and coding human tasks into algorithms that replicate them at scale, at speed, and more accurately.

However, the technology can do exponentially more damage if it goes wrong. While one-off human worker errors can be limited, automation means errors can multiply quickly without oversight, meaning many incorrect transactions occur before the issue is spotted.

Businesses leveraging automation need to be tuned in to the different ways risk can arise. The focus should be put on the design stage of automation deployments, on road-testing processes, and data inputs before live operation. Liability provisions may need to be revisited to ensure it reflects the altered risk profile, e.g. caps based on negative outcomes for the business rather than the number of errors or breaches.