TechCapsules

AI-summarised brief · reviewed before publication

In a recent episode of the ITPro Podcast, Bloomberg’s head of AI strategy and research, Amanda Stent, discussed the persistent challenge of hallucinations in generative AI and how responsible AI practices can mitigate them. Stent explained that while large language models require massive data sets, the quality of that data directly influences output accuracy. She outlined Bloomberg’s internal framework for evaluating data provenance, implementing continuous model monitoring, and enforcing human‑in‑the‑loop verification for critical outputs. The conversation also covered practical steps businesses can take, such as establishing clear governance policies, investing in domain‑specific training data, and fostering cross‑functional AI ethics committees. By sharing Bloomberg’s successes and lessons learned, the episode aimed to provide a roadmap for enterprises seeking to deploy AI responsibly while reducing misinformation risk.