AI drift monitoring with Dominique Shelton Leipzig: silent tests pinging in the background, and accuracy standards taken from whoever did the job by hand.
Dominique Shelton Leipzig, founder and CEO of Global Data Innovation, on AI drift monitoring, the pillar she says most enterprise deployments skip.
Dominique has advised more than 300 companies on AI governance, data ethics, and privacy law, and now works directly with CEOs and boards. She is the author of Trust: Responsible AI, Innovation, Privacy and Data Leadership.
In this conversation with Stephen Sakach and Tullio Siragusa:
• Why AI is never set it and forget it
• Uninterrupted testing as a silent check running in the background
• The window sensor analogy: the accurate state is everything closed
• Getting your standards from the expert who used to do the task
• Supervising humans, because hallucinations are not aberrations
She is equally clear on the consequence of skipping it, which is that without technical documentation nobody can tell when the model started to drift, let alone diagnose why.
Listen to the full episode: https://www.theblisspodcast.com/building-trust-with-responsible-ai-with-dominique-shelton-leipzig/
More from Dominique Shelton Leipzig: https://www.theblisspodcast.com/guests/dominique-shelton-le/
More from the series: https://www.theblisspodcast.com/categories/bliss-businesses-series/