‘These models will always hallucinate’: Seth Dobrin on LLMsDek

Seth Dobrin, founder of Qantm AI and former global chief AI officer at IBM, addresses the complications of large language models (LLMs) in financial services. He highlights the alarming rate of hallucinations in these models and discusses mitigation strategies like retrieval augmented generation. Dobrin also emphasizes the ethical ramifications of data usage, the environmental impact of LLMs, and the need for a responsible framework to harness their potential effectively. His insights shed light on the critical balance between innovation and accountability in AI.