The Road to Accountable AI

Latest Episodes:

Lawrence Snapp (TrustScale): Fighting AI Hallucination Oct 29, 2026

Show notes

Kevin Werbach speaks with Lawrence Snapp, CEO of TrustScale, about why hallucinations remain a stubborn problem and what it takes to verify AI outputs against evidence. Snapp, a CPA by training, compares the current moment to Bill Gates's 2002 Trustworthy Computing memo, which helped launch the modern cybersecurity industry. He argues trustworthy AI is on the same curve, only faster. He is skeptical of labs' self-reported hallucination benchmarks, likening them to a race among mules rather than thoroughbreds, and argues that controls shouldn't be built by the companies being checked.

Most questions pose dto AI have no single factual answer, so TrustScale's goal is to give users complete and accurate information, as with the multiple versions of a monthly unemployment figure, rather than declare what's true. That shift, he suggests, means future AI audit reports will look quite different from financial audits built around materiality.

Snapp walks through Argus, TrustScale's verification tool, which checks AI claims against outside evidence and flags them red, yellow, or green. Argus deliberately stops short of rewriting outputs, because Snapp wants users to pause and avoid what he calls cognitive surrender. He frames hallucinations as cybersecurity threat vectors, and warns that small errors early in an agentic chain can cascade into the kind of large failures that dominate discussions of catastrophic risk.

Lawrence Snapp is CEO of TrustScale, an AI training, evaluation, and assurance company that grew out of e2f, a language data firm founded more than two decades ago. He previously served as CEO of Bryj Technologies, a mobile engagement company; President and CEO of Karmia; Chief Digital Officer and SVP of Corporate Development at National CineMedia; Principal Strategist at Microsoft; and President and COO of Clipper Corporation.

Joshua Tan (Current AI): The Honda Civic Model of AI Oct 22, 2026

Show notes

Kevin Werbach speaks with Joshua Tan, CTO of Current AI, about the case for public AI: The idea that AI can and should be public infrastructure, like highways, libraries, or the internet. Tan traces his path from theoretical machine learning at Oxford to starting the reseaerch group Metagov. He explains what work on decentralized and collective organizations can offer AI governance. He also discusses autonomous organizations run by AI agents, the safety questions that arise when agents hold bank accounts and manage human employees, and why incentives shift once these experiments reach real scale.

Tan positions public AI as a third way between closed proprietary AI and open-weight models, which are struggling to keep pace with frontier-scale capital costs. Current AI focuses on everything beyond the model, aiming to build a fully open source stack on top of fully open models, a Linux for AI. He describes this as a Honda Civic strategy: not at the frontier, but close enough, cheaper, and reliable. The conversation also covers the role of government funding, a new Swiss cooperative, a community-owned data center effort in Maine, and what it would take for Europe and other middle powers to compete.

Joshua Tan is CTO at Current AI. He also leads product and strategy for the public AI movement and leads Stanford AO, an incubator and lab for autonomous organizations at Stanford OpenLab. He previously co-founded Metagov, a lab for engineering institutions, completed his PhD and postdoc in computer science at Oxford, and held fellowships at Stanford, Princeton, and MIT. He founded and led DAOstar.

Transcript

Joshua Tan research page

Current AI

Not just the model (Artifacts Interview)

Brett Levenson (Moonbounce): The Best AI Security System is the One People Will Use Oct 15, 2026

Show notes

Kevin Werbach speaks with Brett Levenson, co-founder and CEO of Moonbounce, about building real-time controls that make AI systems follow the policies companies write for them. Drawing on his years leading integrity work at Meta after the Cambridge Analytica scandal, Levenson explains why content moderation policies there could not be enforced consistently. His team shifted from a decision-first approach to what he calls info-first enforcement, breaking complex policies into simple yes-or-no questions organized in a decision tree. Reviewer consistency jumped dramatically. That atomized approach now underlies Moonbounce, which acts as a semantic firewall that evaluates every input, output, and action an AI takes.

Levenson argues that AI made the application layer non-deterministic, breaking security models that assume predictable software. He describes the challenge of making decisions within a few hundred milliseconds, and a decision memory system that keeps policies adaptive by incorporating customer corrections without hidden state. He also explains Moonbounce's move beyond simple allow-or-block choices to escalating to humans and rewriting outputs. Werbach and Levenson discuss the legal tension around audit trails, payment providers acting as de facto regulators, and why Levenson believes the industry needs measurable, falsifiable standards.

Brett Levenson is co-founder and CEO of Moonbounce. He left Apple in 2019 to lead business integrity at Facebook during the Cambridge Analytica fallout, and later headed Meta's Integrity unit. He co-founded Moonbounce with CTO Ash Bhardwaj, a former Apple engineering leader who built large-scale cloud and AI infrastructure. Moonbounce has raised $12 million in funding co-led by Amplify Partners and StepStone Group.

Transcript

The Facebook insider building content moderation for the AI era

Sharon Goldman (Ground Level AI): AI as It Meets the Real World Oct 08, 2026

Show notes

Kevin Werbach speaks with AI journalist Sharon Goldman about her new venture, Ground Level AI. Goldman describes taking on the AI beat at VentureBeat in 2022, months before ChatGPT launched, and then learning to translate technical developments for Fortune's broad business audience. She explains why she left to launch her own newsletter and podcast, focused on infrastructure, security, governance, policy, and the communities absorbing AI's effects.

Goldman shares her reporting from data center sites, and weighs how much of the data center backlash reflects local impacts versus broader anxieties about AI. The conversation turns to cybersecurity, where the speed and scale of AI-enabled attacks may outpace defenders. Goldman argues that the AI safety and cybersecurity communities, long divided by culture, must now work together. She closes on what she thinks is underappreciated: how quickly things are changing.

Sharon Goldman is the founder of Ground Level AI. She spent four years covering the AI boom, most recently at Fortune and before that at VentureBeat, and has more than 15 years of experience covering technology and business for publications including Business Insider, Forbes, and CIO. She launched Ground Level AI in June 2026 as a newsletter and podcast on Substack.

Transcript

Ground Level AI

Who gets a seat at the table to decide if frontier AI is safe?

AI safety and cybersecurity have long been two different worlds. In the age of AI agents, they're colliding

Robert Levitan (Ethical Tech Project): Unlimited Intelligence, But Where's the Wisdom? Oct 01, 2026

Show notes

Kevin Werbach speaks with Robert Levitan, Chair of the Ethical Tech Project, about what the early internet can teach companies deploying AI today. Levitan, who helped build iVillage and Flooz.com in the dot-com era, argues that the central lesson of those years is that new technologies always carry unintended consequences. Moving fast and breaking things is untenable, he says, now that AI sits on top of the internet and mobile. Levitan explains his essay comparing a coming AI disaster to the 1911 Triangle Shirtwaist Factory fire, which spurred workplace safety laws and helped shape the New Deal. He expects a comparable AI catastrophe to trigger regulation, and urges executives not to wait for it. Trust is a competitive advantage, he argues, and litigation will arrive much faster than it did for the internet. The conversation closes with why Levitan remains an optimist despite the risks, pointing to healthcare's ethical framework and guardrails as a model for other industries.

Disclosure: Kevin Werbach is an advisor to the Ethical Tech Project.

Over thirty years, Robert Levitan has founded, scaled, and advised companies including iVillage, Flooz.com, and Pando Networks, and his recent work focuses on the societal impacts of AI. He studied history and public policy in college and describes himself as a serial entrepreneur whose ventures have ranged from a high-profile IPO to a high-profile bankruptcy to a private sale to Microsoft.

Transcript

Our AI Triangle Shirtwaist Moment Is Coming. Don't Be the Factory Owner (LinkedIn)

Pulse 2.0 interview with Levitan

Harish Peri (Okta): When the Thing Accessing Your Systems Has a Brain Jun 25, 2026

Show notes

In the final episode of our series on governing AI agents, Kevin Werbach speaks with Harish Peri, SVP and General Manager for AI Security at Okta. Peri frames agent governance as the natural next chapter of what Okta has done for two decades: standing in the middle of people accessing technology. The twist is that the new "software" is a non-deterministic agent with a brain, which imposes a much higher security bar. He argues that agents live at the application layer, where the real question is one of authorization: is this agent allowed to take this action or access this data, at this moment, on behalf of this user, given all available signals? Much of the conversation explores why a neutral, independent control plane separate from the frontier models and agent runtimes matters from a cybersecurity standpoint, spreading risk across multiple layers rather than concentrating it in one place.

Peri notes that while awareness of rogue AI is universal, roughly 20% of agents carry about 80% of the risk. He distinguishes security threats like prompt injection and poisoned skill files from "intent mismatch," where an under-specified instruction such as "clean this up" gets read as "delete," and explains how coarse-grained limits, fine-grained context-based authorization, and selectively applied human-in-the-loop checks each play a role in agent governance.

Harish Peri is the SVP and General Manager for AI Security at Okta, where he leads product, go-to-market, and commercial strategy for securing agentic AI. He has more than 20 years of experience across engineering, product management, marketing, and general management, spanning financial services, technology, and human capital management, with prior roles at Salesforce, ADP, and Proxyclick. He holds an MBA from the Haas School of Business at UC Berkeley.

Transcript The Future of AI Security: The Right Architecture for Agents

Secure Your Business Against AI Agents

Logan Kelly (Waxell): The Accidental Agent Governance Company Jun 18, 2026

Show notes

Logan Kelly never set out to build an AI governance solution. Waxell spun out of CallSine, an AI-native sales engagement platform, when the team realized that agents that could act on their own produced a cascade of problems: burning through tokens, accessing databases, creating data-quality issues, and generally doing things no one had explicitly approved. Unable to find existing tooling that addressed the problems effectively, the team built a control plane for agents, which became the foundation of Waxell. In this episode in our series on governing AI agents, CEO Logan Kelly emphasizes that governance should be legible to finance, legal, and compliance teams, not just developers. As he explains, agent governance is less about exotic AI risks than about visibility and control over things companies already care about, such as cost, data access, and who's allowed to do what. Kelly makes the case that the worst outcome isn't an agent misbehaving but companies losing trust in agents altogether and missing their value—arguing that every major technology, from cloud computing onward, arrived with new risks that good governance ultimately made manageable. Looking ahead, he bets that flexible, general-purpose governance will win out over locked-in "walled garden" platforms, as the pace of change in AI keeps accelerating..

Logan Kelly is the Co-Founder and CEO of Waxell, an AI governance and agent observability company that provides a control plane for operating autonomous AI agents in production. The company entered public beta in early 2026. Transcript

Nadav Cornberg (Eve Security): Interrogating Agents Before They Act Jun 11, 2026

Show notes

Kevin Werbach speaks with Nadav Cornberg, co-founder and CEO of Eve Security, about securing agentic AI where it counts: at the moment an agent actually does something. He recounts how customers upended his own assumptions that AI agent security should focus on visibility and after-the-fact detection. Buyers insisted on runtime enforcement first, reasoning that learning a production database was deleted after the fact helps no one. With Eve's "interrogation" approach, when an agent attempts an anomalous, high-risk action, Eve's agent-in-the-loop pauses and questions it about its intent, before approving, blocking, or escalating to a human.

Cornberg describes building a deterministic enforcement layer on top of inherently non-deterministic models, with the system minting explicit rules from observed behavior so that the large majority of everyday requests resolve deterministically. Ultimately, the consequenes are the same whether an unintended action originates in a prompt injection or a simple hallucination. On the perennial human-in-the-loop question, Cornberg argues that escalating everything would drown security teams as the agentic workforce scales, so the platform automatically handles lower-risk cases with justification and reserves human review for the genuinely critical. He closes predicting that "agentic security" will fragment into distinct segments much as endpoint, network, and cloud security once did, and that intent is fast becoming the field's organizing idea.

Nadav Cornberg is the co-founder and CEO of Eve Security, an Austin-based agentic AI observability and policy-enforcement company whose platform governs how AI agents interact with an organization's most critical systems. He brings roughly two decades in product development and engineering, including an early decade in cybersecurity at RSA and Check Point and later work in physical access security across gaming and hospitality before returning to security for the agentic era.

Transcript Founders' blog: Why We Started Eve Security

Venkat Siva (Compfly): Governing Agents at the Execution Boundary Jun 04, 2026

Show notes

Kevin Werbach speaks with Venkat Siva, co-founder and CEO of CompFly AI, about why governing autonomous agents requires a fundamentally different approach than securing traditional software. Siva argues that agents create a genuinely new control problem. Because they decide at runtime which tools to call and which actions to take, governance cannot simply be bolted onto existing MLOps or security platforms built for fixed, deterministic workflows. Instead, control has to move to the "execution boundary" — the point where an agent's decision turns into a real-world action. And agent safety is much more than just model safety.

In practical terms, Siva makes the case for giving every enterprise agent a distinct, cryptographically verifiable identity using decentralized identifiers (DIDs) and verifiable credentials. He addresses the growing problem of "shadow agents," pointing to employees experimenting with powerful open-source autonomous tools inside enterprises, and explains discovery techniques like intercepting traffic to model APIs and watching for who requests LLM keys. He offers the concept of an "autonomy budget": classify actions by reversibility and financial, regulatory, and customer impact, so an agent might autonomously issue a small refund but require human approval for a large one. Drawing on his time at the electric automaker Rivian, Siva closes by contrasting recoverable digital failures with the irreversible stakes of agents embedded in physical systems, arguing that governance there must borrow from safety engineering.

Venkat Siva is the co-founder and CEO of CompFly AI, an early-stage company building a control plane to discover, validate, secure, and govern autonomous agents from code to production. Before founding CompFly with Anand Salodkar, he spent more than two decades building enterprise platform products that help organizations adopt new technology safely and at scale, including work at the electric vehicle maker Rivian. Transcript

The Architecture of Trust (Compfly Manifesto) CoSAI Model Context Protocol Security white paper

Munmun De Choudhury (Georgia Tech): Conversational AI and Mental Health May 28, 2026

Show notes

Conversational AI is increasingly being used as a source of emotional support, even though general-purpose chatbots were never designed for that purpose. Concerns about AI's mental health impact, up to and including suicides, have moved onto the public policy agenda. Munmun De Choudhury, who has been studying the intersection of digital technology and mental health longer than almost anyone, walks through what researchers know, what they don't, and why the answers keep moving.

The conversation centers on the difficulty of governing technologies whose capabilities and patterns of use are both changing every few weeks. De Choudhury invokes the cautionary tale of Google Flu Trends as a warning: any framework that assumes user behavior is fixed will eventually break. She argues that the harms and benefits of conversational AI are not just person-dependent but task-dependent, which makes general-purpose chatbots fundamentally harder to evaluate than the narrow medical AI systems researchers built for decades. She lays out a multi-stakeholder agenda to address AI's mental health risks, and argues that foundation models need to take into account principles from psychotherapy.

Dr. Munmun De Choudhury is the J.Z. Liang Professor in the School of Interactive Computing at Georgia Tech, where she founded and directs the Social Dynamics and Wellbeing Lab (SocWeB). She is one of the most cited researchers in digital mental health and is widely credited with pioneering the computational use of social media data to study mental health outcomes. She co-leads the Patient-Centered Care Delivery research pillar at the Children's Healthcare of Atlanta Pediatric Technology Center, serves on the advisory board for the Australian government's eSafety panel, and was inducted into the SIGCHI Academy in 2024. Her honors include the 2023 SIGCHI Societal Impact Award and the 2021 ACM-W Rising Star Award.

Transcript

Benefits and Harms of Large Language Models in Digital Mental Health

From Lived Experience to Insight: Unpacking the Psychological Risks of Using AI Conversational Agents

Our TOPPODCAST Picks

Follow Us

Stay Connected

The Road to Accountable AI

Related Podcasts

Links

Stay Connected