Three quarters of financial services firms in the UK are currently utilizing artificial intelligence, yet nearly half of these organizations report having only a “partial understanding” of the AI technologies they are implementing, according to a survey conducted by the Bank of England and the FCA.
The survey, which included 118 firms across the financial services sector, indicates that 75% are already using AI, with an additional 10% planning to adopt it within the next three years. This marks an increase from 58% and 14% in 2022, respectively.
A third of all AI applications are third-party implementations, up from 17% the previous year. This reliance on external providers contributes to the gap in understanding: 46% of respondents indicate they have only a ‘partial understanding’ of AI, while 34% believe they possess a ‘complete understanding.’
The findings also reveal that 55% of AI use cases involve some level of automated decision-making, with 24% being semi-autonomous, requiring human oversight for critical or ambiguous decisions. Only 2% of use cases feature fully autonomous decision-making.
Moreover, nearly two-thirds of all AI applications are classified as low ‘materiality’ by their users, with only 16% rated as high materiality.
Respondents highlighted that the primary advantages of using AI lie in generating data and analytical insights, enhancing anti-money laundering efforts, combating fraud, and improving cybersecurity.
In terms of risks, concerns surrounding data privacy and protection, data quality, data security, as well as data bias and representativeness are significant. The risks anticipated to increase most in the next three years include third-party dependencies, model complexity, and the presence of embedded or ‘hidden’ models.
Lastly, the main regulatory constraints perceived in the use of AI include data protection and privacy concerns, followed by issues related to resilience, cybersecurity, third-party regulations, and the FCA’s Consumer Duty.