Cybersecurity Research

OpenAI adds computer-use agents and healthcare plugins to ChatGPT for Healthcare

OpenAI · 16 Nov 2026
Key Takeaway Before letting any AI agent or plugin work inside business software, confirm what data it can access and what actions it can take.

OpenAI has announced new features for ChatGPT for Healthcare, its offering for health organisations. The update introduces computer-use agents, improved clinical answers, and new healthcare plugins.

According to OpenAI, the plugins are designed to work across healthcare software so the tools can be used to get real work done. The announcement does not say here which systems the plugins connect to or how the agents operate within them.

The summary also does not describe the security controls, data handling or access settings that apply to these agents and plugins. Organisations considering the features will need to look to OpenAI's full documentation for those details.

Why it matters

Agents that operate across healthcare software extend an AI tool from answering questions to acting inside systems that may hold sensitive patient information. That shifts the decision from approving a chat tool to approving delegated access, which needs its own review of permissions and oversight. The summary gives no detail on controls, so security leaders should not assume safeguards exist until they have been confirmed.

What to do

  • Request OpenAI's documentation on permissions and data handling for the new agents and plugins before any rollout.
  • Limit any pilot to a small group and the minimum software access needed.
  • Review which healthcare systems the plugins would connect to and who approves that access.
Running AI agents? Eight controls for securing them, mapped to the ISM ->

Summarised by CISO AI from OpenAI, written with Claude Sonnet 5.5. We link back to every original so you can read it yourself.