UK Trends and Developments Contributed by: Sarah Keeling, Piers de Wilde, Dmitry Sachkov and Alexander Greenall, StoneTurn
million for breaching its prior obligation not to onboard high-risk customers while addressing the weaknesses and gaps in its financial crime control framework identified by the FCA in 2020. Despite this, CBPL apparently onboarded and/ or provided services to 13,416 high-risk custom - ers in 2020-23. Artificial Intelligence-Powered Fraud on the Rise Artificial intelligence (AI) is quickly becoming one of fraudsters’ favourite technologies. They exploit advances in especially generative AI – capable of mimicking voices, writing convincing - ly human text, and creating realistic images and video (also known as deepfakes) – to deceive individuals and businesses at scale. In November 2024, the UK bank NatWest pub - lished a list of the ten fastest-growing scams, based on a survey of about 1,500 respondents. Two of them were powered by AI: voice clon - ing scams (third place) and deep-fake celebrity endorsement scams (ninth place). AI is also becoming increasingly popular in more sophisticated and bespoke frauds targeting cor - porations. For example, in early 2024, criminals used AI-powered deepfake technology to imper - sonate senior managers of the UK engineering firm Arup on a video call with the firm’s employ - ee, convincing him to transfer USD25 million to criminals. In what was a staged attack, they reportedly first sent a message to the employee which was purported to be from Arup’s UK- based CFO, summoning him to attend a video call to discuss “confidential transaction” . Pre - sent at the call were the digitally cloned CFO and several other firm staff members, who convinced the employee to complete a total of 15 transfers
to five bank accounts in Hong Kong before he discovered the scam. The sophistication of such AI-enabled fraud makes it highly challenging for potential vic - tims to protect themselves by relying on human vigilance alone. AI should also be at the heart of fraud prevention systems, enabling them to detect deepfake images and voice and spot vari - ous markers and anomalies pertinent to other AI-powered attacks. In May 2024, the UK government launched the Deepfake Detection Challenge, a platform bring - ing together government, academia and indus - try experts to develop solutions focused on the detection of fake media. Following a review of the submissions, the solutions from the consul - tancy Frazer-Nash, the R&D firm Oxford Wave, the University of Southampton and the software developer Naimuri were selected to undergo benchmark testing and user trials. There is no dedicated “AI fraud law” in the UK, as the Fraud Act 2006 covers fraudulent actions regardless of the technology being used. The Online Safety Act 2023 further includes provi - sions aimed at combating AI misuse by target - ing deepfake pornography and synthetic media when used to harass individuals. In more recent developments, on 4 March 2025, the Artificial Intelligence (Regulation) Bill was tabled at the House of Lords. The bill is aimed at regulating AI more generally, and declares “safety, security and robustness” as the first AI regulatory principle. The bill proposes establish - ing a dedicated government watchdog, the AI Authority, to ensure compliance with new legal requirements.
391 CHAMBERS.COM
Powered by FlippingBook