The ITPro Podcast

Amanda Stent delivers substantive, layered information about responsible AI implementation with multiple novel framings: distinguishing responsible/safe/trustworthy; explaining Bloomberg's specific attribution-tracing methodology; and discussing the RAG limitations paper. However, roughly 30% of the episode consists of setup, acknowledgments, and host throat-clearing that dilutes insight density. The financial services context-setting, while relevant, occasionally edges toward expected background rather than novel claims.

“we check that every statement that's being produced by one of our generative AI systems can be traced back to either a statement in a source document or to a data element or set of data elements in a database or the result of a calculation”

“we have a paper from last year showing that RAG systems are not necessarily safer in the sense of less likely to hallucinate than non RAG systems”

Stent avoids obvious platitudes and presents several genuinely fresh angles: the distinction between efficiency-focused AI (traditional ML in trading) versus effectiveness-focused AI (new analyst workflows), neurosymbolic AI linked to ontologies, and the principles-based UK regulatory model as a model for responsible AI. However, the core frameworks (RAG, guardrails, attribution-checking) are increasingly standard in responsible AI discussions, and the guest doesn't challenge dominant narratives or offer strong contrarian positions.

“I think it's about effectiveness. So it's helping people become more effective in how they use AI”

“principles based, technology agnostic approach”

Amanda Stent is exceptionally well-positioned: Head of AI Strategy and Research in the office of the CTO at Bloomberg, a 15+ year practitioner at a company with deep AI implementation history since 2009. She speaks from direct responsibility for responsible AI deployment at a regulated, high-stakes financial institution where hallucinations carry real market consequences. She has published relevant whitepapers and brings both operational authority and research credibility.

“I'm Amanda Stent, Head of AI Strategy and Research in the office of the CTO at Bloomberg”

“Bloomberg has been using AI since about 2009”

Stent provides concrete details about Bloomberg's tech stack (17,000 news providers, 1,000+ research brokers, 400M+ documents, agentic AI systems, guardrails) and specific use cases (analysts saving 'days and days and days' of time, portfolio managers writing on-demand reports, FCA testing framework). However, she avoids quantified ROI, specific timelines, and named examples of hallucinations that were caught or costs avoided. The financial services sector discussion relies on historical narrative rather than current metrics.

“The terminal gives users access to more than 17,000 news providers, not just Bloomberg News, more than a thousand research brokers, more than 400 million documents from companies themselves, and billions and billions and billions of ticks”

“one of them said days and days and days, save days and days and days of time to help them write their research reports more quickly”

The hosts ask substantive opening questions and follow up meaningfully on Bloomberg's use of AI and regulation. However, they miss opportunities to push back, challenge, or go deeper. When Stent claims RAG isn't necessarily safer, no host probes what Bloomberg uses instead or why. When she discusses job displacement, no one presses on evidence or downside scenarios. Most follow-ups are gentle extensions rather than sharp interrogations. The conversation is friendly but lacks productive friction.

“Is responsible AI trustworthy AI a way to get around these false but confident answers that generative AI in particular can put out?”

“Is it that you have AI cross checking other AIs work or how much do you have a human involved in all of this?”