UK Bank employees are right to be anxious about AI compliance.
Recent research suggests AI adoption continues to be incentivised in financial workplaces without correct workflow implementation, risking potential compliance breaches. Grace McNicholas asks who’s responsible if banks don’t use AI securely.
85% of British bank employees fear AI usage will lead to compliance issues, but many leaders continue to mandate its use.
New research by Camunda surveyed 1,000 decision makers involved in automation and 5,000 UK bank employees who use AI in their day-to-day work. The survey reveals ongoing concern about AI usage in banking and finance – mainly from employees.
98% expect process challenges to contribute to future problems.
Camunda found users are anxious about being accountable for AI workflows they weren’t consulted on, with 69% of employees claiming there had been no consultation prior to AI implementations. Furthermore, 46% said they would consider changing jobs over poorly implemented AI mandates, and 65% have said they’d use AI more effectively with different implementation patterns.
Strange, then, that 89% of leaders claim AI makes their teams more productive.
Such employee fears are far from irrational: at a recent closed-door roundtable, several firms described compliance issues and disciplinary actions taken relating to AI usage – such as client-sensitive information being given to public AI tools.
Unless institutions establish proper workflows including measures to mitigate risk when using AI, business leaders can’t blame their employees for being concerned. Some may claim employee fears are exaggerated – after all, less than half (40%) of the organisations surveyed said they’d experienced an AI compliance problem in the last year. To which the cynic might respond that there’s insufficient regulation around AI at present, and many incidents go unreported, as we discussed in a previous article on this topic
Employee concerns will persist until safeguards emerge.
The good news is that one firm at the recent closed-doors roundtable found new restrictions on AI usage have been successful in controlling risks. This company has limited employee AI usage to a single model that sits within its own secure data environment – a good example of practical action on AI safety which other firms would do well to follow.
While organisations continue to mandate AI use without sufficient security measures, liability for any breach of data confidentiality should sit with them, and employee concerns about compliance breaches are legitimate. As AI use soars and regulation struggles to keep pace, banks should be asking who pays the price for their headlong rush towards Large Language Models?