Accuracy concerns block AI adoption in UK finance, Bloomberg survey finds
Half of UK financial leaders cite inaccurate outputs as their biggest barrier to adopting AI, according to a Bloomberg poll of over 100 senior decision-makers at its AI in Finance Summit in London on April 16.
Hallucinated facts and numerical errors topped the list of concerns. A further 27% pointed to lack of explainability, meaning accuracy and transparency together dominate how finance professionals evaluate AI tools.
The survey reveals a clear pattern in what builds trust. When asked what gives them confidence in AI systems, respondents chose verifiable features: 32% selected source attribution, 30% highlighted built-in error checking, and 25% chose human oversight. Only 9% said sophisticated language and reasoning mattered most, and 5% prioritized speed.
This preference for control over capability reshapes how vendors should build financial AI. Amanda Stent, Head of AI Strategy & Research at Bloomberg, said: "Trustworthiness depends on whether an AI's outputs can be interrogated and validated. Solving this challenge depends on attribution, transparency and the quality of the underlying data so outputs can be traced to their sources, validated for accuracy, and confidently used in decision-making."
Despite accuracy concerns, appetite for advanced applications is strong. Nearly two-thirds of respondents identified full workflow AI assistants as the most exciting next development, far ahead of personalized portfolio insights (9%) or no-code quant tools (12%).
The gap between these findings matters. Finance leaders want end-to-end AI integration-but only where systems deliver reliable, verifiable outputs at scale.
Bloomberg announced a 2026 roadmap for ASKB, its conversational AI interface currently in beta. The tool is designed to augment investment workflows using Bloomberg's generative AI and LLM technology grounded in trusted data, with accuracy and control built in from the start.
For finance professionals evaluating AI for finance, the survey underscores a single message: verification matters more than capability.
Your membership also unlocks:
