Topic starter
25/10/2026 11:09 am
TL;DR: OpenAI’s Dots shift ChatGPT agent security from supervised sessions to always-on agents with their own cloud computers, credentials and app connections, according to NeuralTrust, while OpenAI still says safeguards do not eliminate prompt injection. That makes access approval, logging and scoped control the governance baseline, not an optional hardening layer.
Editorial analysis by NHI Mgmt Group, based on content published by NeuralTrust: “ChatGPT Agent Security: What Dots Change”.
By the numbers:
- Dots launched on September 29, 2026.
- Dots can readily connect to over 4,000 apps.
- GPT-6 Astra scored an 8.5% attack success rate on Gray Swan's IPI Arena.
Key questions
Q: What breaks when AI agents are treated like standard human users?
A: You lose visibility into effective permissions, expected behaviour, and real blast radius.
Q: Why do AI agents increase access risk even when they are visible?
A: Because visibility does not reduce privilege on its own.
Q: How should security teams handle agent-to-agent delegation in approval workflows?
A: Treat it as a separate trust boundary, not as a convenient relay.
Practitioner guidance
- Define a standing identity model for every agent Assign an owner, a lifecycle state and a clear offboarding process before enabling persistent access or memory.
- Restrict connector scope before first use Limit each dot to the minimum set of apps and data paths required for the pilot, then review every connector as privileged access.
- Map high-risk actions to explicit approval gates Require human confirmation for password changes, transfers, deletions and any new security-sensitive access, even when the platform allows automation.
Bottom line: OpenAI’s Dots shift agent security from a supervised chat model to a standing identity with persistent access, memory and connectors.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →