Dr. Freddie Seba

AI Governance Keynote Speaker  ·  Author  ·  Scholar Operator

EdD · USF  ·  MBA · Yale  ·  MA · Stanford  · Teaching · UIC  ·  20+ years · Silicon Valley Founder & Global Executive · Digital Health · Fintech · Higher Ed.

Issue #58: Governing the Agentic Era Before Drift: Insights from an AI Keynote Speaker

The Seba 12 Ps for Boards, Health AI, Wearables, Sovereignty Claims, and the New “Filtering Problem”

AI Ethics & Governance for Leaders, Boards & Trustees

By Dr. Freddie Seba

© 2026 Freddie Seba. All rights reserved.

Executive Reflection: Autonomy Is Not a Feature — It’s an AI Governance Variable

This week’s signal is not “AI is getting smarter.”

It’s that agent autonomy is becoming measurable — and therefore governable.

@Anthropic published real-world evidence from millions of human-agent interactions showing autonomy is not a single switch; it shifts with user behavior, product design, and oversight choices. In the longest-running @Claude Code sessions, autonomous runtime increased materially over a short period, and experienced users increasingly enabled auto-approve while still interrupting when needed. The message is clear: effective oversight requires post-deployment monitoring infrastructure and interaction patterns that balance autonomy and risk.

That’s the AI governance question for boards and executive teams:

Are we treating autonomy as a capability claim — or as a controllable deployment choice with clear decision rights, monitoring, and escalation?

What We’re Seeing (Signals)

1) Autonomy is shifting from a story to an AI governance measurement discipline

@Anthropic’s autonomy research lands for governance leaders: the same model can be deployed with radically different autonomy depending on the interaction paradigm and oversight scaffolding. Their conclusion points directly to what boards should require: autonomy is not static, and it must be monitored and governed after deployment.

AI governance translation: require an “autonomy profile” per AI use case—what the system can do, what it’s allowed to do, what it may do by default, and what requires explicit approval.

2) Clinical AI maturity is moving from “model building” to “deployment craft.”

A signal I love seeing: training materials that teach operationalization as a first-class skill.

@MIT OpenCourseWare published HST.953 (“Clinical Data Learning, Visualization, and Deployments”), a graduate course explicitly focused on operationalizing ML in healthcare—robust/private/fair ML, visualization for clinical utility, and the implementation science needed to tie models to real clinical use.

AI governance translation: clinical AI isn’t a data science project. It’s an operational system that needs monitoring, workflow integration, and accountable ownership.

3) The doctor–patient relationship is now part of the AI governance surface

A 2026 narrative review in Healthcare (MDPI) examines how AI is reshaping the doctor–patient relationship—reminding us that governance isn’t only technical; it’s relational. (Authors include @Marco Casciaro, @Sebastiano Gangemi, and @Gabriella Martino.)

AI governance translation: trust isn’t a downstream effect. It’s an input requirement. If AI changes how patients interpret, question, and comply, your oversight model must include that human layer.

4) Patient-facing AI is improving readability — and forcing a new disclosure standard

A systematic review in @The Lancet Digital Health suggests that LLM-simplified radiology reports can improve readability and patient-perceived understanding while maintaining a measurable error profile.

Pair that with a concrete patient-trust signal: @YaleNews highlights how consumers use chatbots for health advice and why leaders should be cautious about over-reliance, especially when outputs harden beliefs or are treated as clinical truth.

AI governance translation: once outputs become patient-facing, “helpful” becomes a regulated interface. You need disclosure baselines, monitoring, and error-escalation pathways.

5) Healthcare operators are converging on “practical AI.”

A useful practitioner signal: Practical AI in Healthcare (S1E24) features @Bob Wachter (UCSF Chair of Medicine) discussing what matters in real deployment—why AI doesn’t need to be perfect, but does need to be governed, implemented responsibly, and evaluated against systems already failing at scale.

AI governance translation: the benchmark is not model performance in isolation. The benchmark is system performance under accountable deployment.

6) Wearables are becoming autonomous distribution points

Reports indicate @Meta is reviving smartwatch plans for 2026 with health tracking and an AI assistant—expanding the surface area where consumer behavior, health inference, privacy, and secondary data uses collide.

AI governance translation: wearables aren’t “devices.” They’re data + inference ecosystems. Treat them as identity, consent, and secondary-use governance problems.

7) Revenue pressure is now an AI governance risk category

The @Center for Democracy & Technology (CDT) issue brief “Risky Business: Advanced AI Companies’ Race for Revenue” surfaces a governance reality boards can’t ignore: commercial incentives shape deployment choices, safety posture, and disclosure practices across leading labs and platforms—including @Google, @Meta, @Anthropic, @OpenAI, and @xAI.

AI governance translation: when revenue sets the tempo, governance must be the brakes-and-steering system—clear decision rights, monitoring, escalation, and audit-ready evidence.

8) The “filtering problem” is becoming the dominant AI governance problem

Across domains, the same pattern keeps appearing: scale increases variance. More output means more noise. The institutions that benefit are those that build filtering infrastructure—clear standards for evidence, accountability, and reliability in real use.

This is why @Ethan Mollick’s repeated observation matters: it’s still surprising how broadly LLMs work, but that doesn’t remove the need for AI governance—because “it works” is not the same as “it is safe, accountable, and trustworthy at scale.”

Practical guidance like @oneusefulthing’s “which AI to use” framing—and engineering discussions like @alindnbrg’s enterprise agentic AI signal—are pointing to the same AI governance truth: model choice is only one variable; oversight is the system.

The Seba Framework: The 12 Ps of Responsible AI Oversight ©

(How I translate signals into board-ready AI governance.)

  1. Purpose — mission alignment vs. cost extraction
  2. Problems — decision-relevant framing (not metric chasing)
  3. Profits — who benefits vs. who bears risk
  4. People — patients/students/workforce; lived impacts
  5. Planet — compute, infrastructure, scale costs
  6. Process — lifecycle monitoring + incident learning
  7. Policy — risk-specific rules (health, youth, education, employment)
  8. Protections — vulnerable populations + escalation paths
  9. Privacy — enforceable limits on data use/training
  10. Provenance — traceability of data, models, vendors
  11. Preparedness — board competence + governance cadence
  12. Product Ownership — institutions own outcomes once AI acts

Board-Ready Next Step (This Week’s Practical Ask)

If you only do one thing this quarter:

Require an “Autonomy & Accountability Sheet” for every AI use case — one page that states:

  • the autonomy level allowed (what it can do vs. what it may do)
  • the approval model (auto-approve rules, human checkpoints)
  • the monitoring signals (errors, drift, overrides, complaints)
  • the escalation path (who can pause, who must be notified, how fast)

This is the AI governance bridge between “agent capability” and “institutional accountability.”

Podcast Series: Episode 7 (Coming This Week)

Episode 7 of my podcast series drops this week—a companion to this newsletter focused on how practitioners are actually deploying AI (and where agentic systems break) under real operational pressure.

I’m currently editing and producing work with two exciting guests/episodes, so I’m holding the names until the final cut is locked.

AI Governance with Dr. Freddie Seba is available on Spotify, YouTube, Apple Podcasts, and the Substack app.

About the Author

Freddie Seba is a researcher and practitioner—and an AI keynote speaker and AI ethics speaker—focused on AI ethics and AI governance for leaders across higher education, healthcare, and financial services.

He holds an MBA (Yale University), an MA (Stanford University), and an EdD in Organization and Leadership (University of San Francisco), with a dissertation on AI ethics and governance defended in Fall 2025.

He writes AI Ethics & Governance for Leaders, Boards & Trustees and hosts the companion podcast AI Governance with Dr. Freddie Seba, translating practitioner signals into board-ready oversight: decision rights, risk tiering, vendor accountability, monitoring, and incident preparedness.

Speaking, Briefings, and Board Workshops

I keynote on AI governance, risk, trust infrastructure, and institutional legitimacy.

I also deliver AI executive workshops and board briefings that bring strategic framing and practical takeaways for boards and senior leadership—accountability, transparency, safety, responsible adoption in regulated environments, judgment under uncertainty, escalation design, and governance maturity—across business and educational engagements, executive briefings, and board workshops: inventory → tiering → controls → dashboards → incident drills.

To book an AI keynote speaker engagement, AI ethics speaker talk, AI corporate event talk, AI executive briefing, or AI board workshop: connect via freddieseba.com and @LinkedIn.

And please subscribe to the newsletter and follow the podcast (available on Spotify, YouTube, Apple Podcasts, and the Substack app).

Gratitude

Grateful for the communities that nurture this work and keep pushing for trustworthy, accountable AI: University of San Francisco, Stanford HAI, AMIA Informatics, and the Coalition for Health AI.

Transparency + Disclaimer

Educational content only. This newsletter does not constitute legal, medical, clinical, insurance, financial, or professional advice.

Drafted and refined with AI-assisted tools for synthesis and clarity. Final editorial control and responsibility remain with the author.

© 2026 Freddie Seba. All rights reserved.

AIGovernance #ResponsibleAI #BoardOversight #ExecutiveLeadership #AILeadership

AIandTrust #HealthcareAI #AIinEducation #SovereignAI #RiskManagement

DigitalTrust #EnergyAndAI #SiliconAlley #TechPolicy

Sources

#AIKeynoteSpeaker #AIEthicsSpeaker #AIExecutiveWorkshops #AIEthicsBoardWorkshops #AISpeaking #AIworkshops