OpenAI expands ChatGPT Agent to Pro with computer use and web browsing

Read the original article →

What happened

OpenAI expanded ChatGPT Agent to Pro subscribers on July 28 with computer use, web browsing, and multi-step task execution from a single prompt.

The agent can fill forms, book appointments, and draft documents using a virtual browser and desktop environment. It runs on GPT-5.5 with safety classifiers and requires user confirmation before irreversible actions. Enterprise rollout is planned for August.

Why it matters

Computer-use agents are the next step beyond chat and coding assistants. Giving Pro users access tests the feature at scale before enterprise deployment.

Confirmation gates for irreversible actions are a necessary safety default, but they also limit full autonomy.

MintedBrain take

Treat computer-use agents as semi-autonomous helpers, not unsupervised workers. Review confirmation prompts carefully and avoid connecting them to accounts with broad permissions until you understand failure modes.

References

This article was originally published at The Verge. For the full piece, read the original article.

Discussion

  • Loading…

← Back to News