Part 2 builds on the previous episode, describing how to establish AI governance in an organization, and what AI literacy truly means for business results.
Anthony Scriffignano, PhD is one of the world's best data scientists — a thinker who has spent more than 40 years at the intersection of linguistics, advanced algorithms, business intelligence, and the very human problem of making good decisions from imperfect information. Anthony is currently a Distinguished Fellow and member of the Alfred Lee Loomis Innovation Council at The Stimson Center. He also serves on advisory committees in government, the private sector, and academia, and has briefed the National Security Telecommunications Advisory Committee and contributed to reports to the President on Big Data Analytics and Emerging Technologies. He served on the OECD Network of Experts group on implementing Trustworthy AI.
His most recent publication is a co-authored chapter — "Illuminating the Invisible: AI's Role in Understanding Complex, Opaque, and Dynamic Ecosystems" — in the Springer volume Artificial Intelligence Revolution (2026, edited by Max Rangeley and Nicholas Fairfax), co-written with Dr. David Bray.
Key Takeaways:
Compliance is what you must do; governance is what you should do. Governance costs money, time, and opportunity. The right question is not "how do we maximise governance?" but "where do we sit in the continuum between maximum safety and maximum agility, and are we there on purpose?"
The human-in-the-loop section challenges the common assumption that more humans equals better decisions. Anthony identifies three limits: humans cannot always make decisions fast enough; even the same human given the same decision repeatedly will not always make the same call; and humans have bias, fatigue, and optimism. The emerging solution is cognitive AI that watches experts making decisions, converges on patterns, and eventually can substitute a synthetic human into the loop for low-stakes votes.
Agentic AI is described as a system of semi-autonomous entities, each doing a small piece of a problem, orchestrated together, some inside the enterprise and some outside. Auditing every individual agent is not feasible. The answer is a clinical layer that monitors the health of the agentic ecosystem as a whole — taking the pulse, measuring blood pressure — and raises flags when the system-level behaviour deviates from its guardrails. The language to do this does not fully exist yet.
On shadow AI: Anastassia's survey found 64% of respondents in SMEa are using AI tools their employer has not sanctioned. Anthony suspects the real number is higher and points to non-response bias. His prescription is not HR enforcement, but understanding the underserved need: if employees are using shadow AI, the organisation has failed to provide them with the tools they need. Create a safe sandbox, serve the need, and watch what they build.
Chinese LLM approach takes a fundamentally different approach from the US— leaner, faster, cheaper, focused rather than voracious.
AI literacy versus AI fluency. Fluency — knowing the buzzwords, knowing how to prompt — is the lowest form of useful knowledge. Literacy means knowing what you do not know: understanding the limits of the tool, the provenance of the output, and when you should not trust it.
https://www.stimson.org/ppl/anthony-scriffignano/
Anastassia Lauterbach - LinkedIn
First Public Reading, Romy, Roby and the Secrets of Sleep (1/3)
First Public Reading, Romy, Roby and the Secrets of Sleep (2/3)
First Public Reading, Romy, Roby and the Secrets of Sleep (3/3)
