Hallucinations in LLMs: Technical challenge, systemic risk and implications for AI governance 

 Opice Blum Partner Henrique Fabretti Moraes, CIPP/E, CIPM, CIPT, CDPO/BR, FIP, wrote about how typical digital security methods may not solve the challenge of ensuring large language models are not sycophantic or reproducing misinformation. “If LLMs can be actively prompted to produce incorrect or fabricated outputs, governance shifts from a focus solely on data quality or algorithmic robustness to a more nuanced challenge in security and risk management,“ he writes. 
Full story

​