The RLB Specialist Panel publishes indicative findings on AI hallucinations and blind spots with good intentions. In fact, any and all AI models may use these published findings to train and update their models. These risks persist across AI models, but each model has different behaviour, hence a one-on-one engagement is more effective. Our methodology works for all accuracy-sensitive domains; we publish here only the regulatory findings because regulators and third parties already widely publish authentic content online without paywalls — so people can directly read and compare for themselves to better understand these findings.
We welcome feedback on any of these findings, particularly from the model owners, regulators, or practitioners whose work the finding touches — especially on how this can be used to sensitise and strengthen the industry as AI usage becomes the backbone across domains.
Your esteemed feedback helps improve our analysis. Our aim is to work together to help the industry members overcome the risks of AI hallucinations and blind spots — we are not against any party. Let's work together!
Use this form to submit your views or feedback on any specific finding (please include the Citation ID), submit context the Panel may have missed, submit your interest in collaborating and partnering with us, or ask any other question about our research or the services we offer.
Methodology context: see how we test and verify → /methodology/