Experian builds cyber resiliency into its AI infrastructure
Enterprises racing to deploy AI at scale are discovering that models depend on the infrastructure beneath them. As cyber resiliency becomes central to AI deployment, organizations must address foundational issues such as data quality, access controls and recovery. That work is particularly important in data-intensive industries such as financial services. Experian PLC has built its […] The post…
As AI adoption accelerates across industries, organizations face the challenge of ensuring their AI infrastructure remains resilient and secure. Rodrigo Rodrigues, group chief technology officer of Experian PLC, emphasized during the AI Cybersecurity and Resilience Summit that weak infrastructure foundations can create risky conditions when AI is layered on top.
Experian has prioritized AI infrastructure resilience, addressing critical areas such as data quality, access controls, and recovery processes. Rodrigues highlighted that cyber resiliency cannot be assumed solely by the existence of backups, as NIST's AI Risk Management Framework is deeply integrated into Experian's AI pipelines.
To mitigate risks, Experian enforces strict policy validation and automated stops in AI pipelines when controls fail, ensuring no manual overrides. The company also tests AI recovery through scheduled drills and clear recovery objectives. Rodrigues stressed that bad training data can cause models to learn incorrect patterns, and thus, AI must reflect the permissions given to it through least privilege access.
Experian has even developed an internal AI agent marketplace, treating AI agents as versioned assets with defined owners, lifecycle policies, and sunset criteria, to help manage the complexity of AI deployment while maintaining a fast-paced development environment.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
