Building Trust in AI: ‘Trust, Verify’ Approach Needed for Reliable Audits

The increasing sophistication of AI models presents unique challenges to current auditing processes, impacting reliability and fostering a lack of trust in critical sectors like finance. Samuel Pearton, CMO at Polyhedra, calls for a ‘trust, but verify’ approach to address this issue. This means comprehensive reviews throughout the model lifecycle – training, inference, and deployment – to rebuild trust in AI applications.