London finance staff embrace AI but want human oversight

Bloomberg has published a survey showing widespread use of artificial intelligence among London financial services professionals. The poll found that 88% use AI tools daily or weekly.

The research covered 500 full-time employees at London financial firms with more than 50 staff, spanning front-office, middle-office and senior management roles. It depicts a workforce that has already integrated AI into routine work while still questioning where human judgement should remain central.

One of the clearest findings was the gap between productivity gains and overall workload. While 90% of respondents said AI had boosted productivity, nearly three-quarters said it had not reduced the amount of work they were doing.

Rather than cutting hours or materially reducing task volumes, the technology appears to be changing how time is spent. Among those who said AI had freed up time, 51% said they were redirecting it to higher-value activities such as strategy, decision-making and client relationships, while 48% said they had reduced time spent on routine administration. A further 36% said they were collaborating more with colleagues, and 32% said they were finishing work earlier.

Skills shift

The survey also suggests AI literacy is becoming a factor in career progression. Some 85% of respondents said AI skills would soon matter more than traditional skills for their next promotion or career move, while 69% said colleagues who resist developing those skills have no place at their firm.

Workers are not relying solely on formal corporate programmes to build those skills. The average respondent used more than three methods to improve AI knowledge. The most common was using AI tools themselves to learn AI, cited by 53%. Experimenting in personal time followed at 50%, ahead of watching videos and online guides at 44%, learning from colleagues at 42% and formal employer training at 40%.

The findings come as policymakers and employers weigh how quickly financial services staff need retraining as AI systems spread through research, compliance, operations and client-facing work. They also suggest workers in the sector see the shift less as a distant structural change and more as an immediate professional requirement.

Amanda Stent, Head of AI Strategy & Research in the CTO Office at Bloomberg, said the results showed a practical and self-directed approach to learning.

"Becoming effective with AI doesn't require a PhD in computer science - it requires curiosity and domain expertise. What's striking about these findings is the initiative London's financial services professionals are showing in response to the rapid evolution of AI. They're learning by doing, experimenting and asking the tools themselves. That kind of self-directed curiosity is exactly what thriving in an AI-driven industry requires," said Amanda Stent, Head of AI Strategy & Research in the CTO Office at Bloomberg.

Concern rankings

Despite regular use of AI tools, respondents did not rank job losses for junior staff among their top concerns. When asked to choose their three biggest concerns about AI, only 30% selected the loss of entry-level and graduate roles, making it the seventh most common concern out of eight options.

That placed it well below AI-enabled cyberattacks and financial crime, selected by 45% of respondents. The ranking suggests London finance professionals are more concerned about misuse, control and operational risk than the direct displacement of early-career roles.

The survey also points to limits on professionals' willingness to delegate decisions to AI systems. Just over half, or 51%, said risk and compliance was the area where human oversight mattered most, followed by investment and trading decisions at 40%. Only 1% said human oversight was unnecessary in most cases.

Stent said any broad plan for AI adoption in finance needed to reflect the distinction between automation and responsibility.

"Removing the friction from knowledge work doesn't remove responsibility. As AI handles more information retrieval and synthesis, humans shift their time from manual research to acting as the ultimate arbiter. Any government plan for AI adoption in financial services needs to account for that. A model may detect a pattern; it takes a trader or analyst with years of experience to understand whether that pattern is a meaningful signal or just noise," she said.

The emphasis on oversight is notable in a sector where errors can have immediate regulatory, financial and reputational consequences. Even as firms adopt AI in research and workflow management, the results indicate that staff still expect experienced professionals to take responsibility for interpretation and final judgement.

That balance between acceptance and caution runs through the survey. Professionals are using AI frequently, investing their own time in learning it and increasingly linking it to career advancement. Yet they do not appear to see the technology as a substitute for senior judgement in critical financial decisions.

Stent said the findings showed finance remains rooted in human decision-making.

"Finance remains a fundamentally human endeavour, enhanced and empowered by technology. In a world where information is abundant, judgement and leadership are the scarce resources. What this research shows is that London's financial services professionals understand that and are positioning themselves accordingly," she said.