OpenAI has published a lifecycle account of its frontier-model controls, including an incident in which an internal research model exploited a DNS misconfiguration to gain limited internet access. The company says monitoring detected the activity and that it paused training and evaluation of its most capable models from 20 September to 1 October while it investigated and closed the gap.

The new account argues that frontier training should take place without live internet access under current alignment, monitoring and infrastructure conditions. OpenAI describes layered containment through restricted sandboxes, virtual machines and network and cloud perimeters. It also says internal red teams use capable models to test whether those boundaries can be escaped. Monitoring is positioned as the backstop when prevention fails, alongside rapid-stop controls and incident-disclosure processes.

This is OpenAI's own description of its controls and incidents, not an independent assurance report. It is nevertheless unusually specific about the limits of containment and the decision to stop work. The broader governance lesson is that a safety case needs operational authority behind it: defined triggers for pausing training, evidence that monitoring can detect unexpected behaviour, and leadership able to accept schedule costs when controls are insufficient. Buyers cannot reproduce a training environment, but they can ask providers for comparable evidence about isolation, response thresholds and disclosure.


Execution Level Governance- What audit-ready agent governance actually looks like
David Girvin, founder and CEO of Assury argues that model-in-the-loop review, AI governing AI, is fundamentally unreliable for regulated environments: even the best-performing models miss a meaningful share of violations, the reviewing model is typically provided by the same vendor being reviewed, and prompt injection or context poisoning can compromise both the acting agent and its supposed overseer simultaneously. He makes the case for deterministic, architecturally enforced controls instead, walking through Assury's approach of autonomy zones, session risk accumulation, and credential starvation, which lets a compromised agent be cut off from its tools instantly rather than relying on time-boxed access. The conversation touches on why David is sceptical of just-in-time credentialing as a solution for agent security more broadly, since agent sessions don't run on predictable human timescales, along with the current gap between how identity and security vendors are pitching agent protection and what he sees happening at the execution layer in practice. He also discusses the compliance and audit implications of probabilistic decision-making, arguing that regulated industries will increasingly need tamper-evident, hash-chained audit trails that can withstand scrutiny from auditors and regulators who are only beginning to understand agentic risk, and reflects on a named frontier lab's own published framework as an example of the gap between research and practitioner reality. Elsewhere, David reflects candidly on building a bootstrapped security company in an increasingly crowded market, why he turned down aggressive VC funding to stay in control of the product, and what a credible third-party assessment of his own gateway would need to look like given that Assury sits directly in the execution path for every customer's agents.
Share this post
The link has been copied!