A significant increase in the use of AI by firms is revealed in the latest yearly joint survey by the Bank of England (BoE) and FCA. Figures show 75% are already using the technology, with another 10% planning to do so over the next three years. Those figures were 58% and 14% respectively in 2022.

The findings also support anecdotal evidence pointing to the rapid adoption of foundation models, with 17% of use cases engaged in this type of complex machine learning.

The detailed study, looking into the adoption and utilisation of AI in UK financial services, has looked into different aspects of AI integration within the finance industry, including adoption and use, strategies and governance, benefits, risks and constraints.

The percentage of third-party exposure has also doubled, from 17% in 2022 to around a third. The authors believe this will increase further with a decrease in outsourcing costs and increase in the complexity of models.

And 55% of all use cases have some degree of automated decision making capacity, with 24% being semi-autonomous and 2% fully automated.

Materiality remains low though, with 62% of the use cases being categorized as “low materiality” by the firms. Only 16% have been categorized as “high materiality.”

Understanding of AI

The survey shows that understanding the technology, to some extent, depends on how much a firm relies on third-party technology providers for their day-to-day use.

Of all the firms that took part in the survey, 46% said they had “partial understanding” of the technology they use. Another 34% said they had a “complete understanding” of AI.

Authors noted that the degree or percentage of understanding was higher among firms that used internally developed models and relied less on third-party providers.

When it comes to benefits of AI, firms have reported that the technology is most useful in areas such as data and analytical insights, anti-money-laundering (AML) and combating fraud, and cybersecurity.

But the findings have also shown that firms have many concerns in relation to data. Some of the biggest perceived risks have to do with “data privacy and protection, data quality, data security, and data bias and representativeness.”

Also, “cybersecurity is rated as the highest perceived systemic risk both currently and in three years,” the survey has found out.

Data protection and privacy were found to be the largest perceived regulatory constraints to AI adoption. Safety, security and robustness of AI models were found to be the largest perceived non-regulatory constraints.

Data and governance

The survey also looked into how firms dealt with data that was exposed to AI technology, and the frameworks or procedures firms used for internal AI governance.

• 84% of the respondent firms said they had a specific staff member, or members, who was responsible for the AI framework within the organisation.
• 82% of the firms also used an AI framework, principles, guidelines or best practice.
• 79% mentioned data governance.

And when it comes to accountability, 72% of the firms said they allocated it to their executive leadership. This is followed by developers and data science teams (64%), and business area users (57%).

The management of data remains a concern, and findings of the survey show that in most cases this task is not AI-specific in the firms that took part.

Data privacy and security continues to be a priority, with 19% of respondent firms using AI-specific practices and 66% using non-AI-specific practices. 

Comparison with other surveys

When it comes to adoption and utilisation of AI by the financial industry, the findings are similar to those of a recent study by the Financial Stability Boards (FSB).

Among the key findings of that study was the fact that AI adoption in the financial sector is on the rise, and that most use cases revolve around two key areas; enhancing internal operations and improving regulatory compliance.

The FSB survey also revealed that the use of AI to generate new revenue streams in the financial sector is not widely observed, though this could change in the future.

Those findings were also supported by a recent but separate study by the Boston Consulting Group (BCG) which looked into how firms around the world were finding value in AI adoption.

The BCG survey showed that only 4% of the more than 1,000 global firms who took part are utilising AI to find true value, be it in revenue generation or other sectors of their businesses.