OpenAI's ChatGPT Work can now operate on login-required websites. It shows a login screen, and users enter credentials themselves.
The AI never sees usernames or passwords.
Available on Plus, Pro, and Business plans.
What happened
OpenAI announced on August 25 that ChatGPT Work can now perform tasks on websites that require login. Users log in themselves, and the AI continues the work while maintaining the session.
Why it matters
Previously, tasks would stop on sites needing authentication. The new feature lets ChatGPT handle more real-world workflows, like submitting email invoices to accounting software or saving apartment listings, without the AI ever seeing usernames or passwords.
What to watch
The feature is available for Plus, Pro, and Business plans, and works on web and mobile versions. For critical actions like confirming reservations or payments, ChatGPT will always ask for confirmation.
Ask the AI about this article →
OpenAI's announcement marks a step forward for AI agents that can perform tasks across the web. By allowing ChatGPT Work to operate on login-required sites, the company addresses a common limitation where tasks would stop when authentication was needed. The approach of having users log in themselves, rather than storing credentials, aims to balance convenience with security.
The company emphasizes that usernames and passwords are not sent to the model and are not used for training. Admins can manage which sites ChatGPT Work can access, and the feature includes a confirmation step before critical actions like payments. This suggests a cautious rollout that prioritizes user control and safety.
The use cases outlined, such as submitting invoices to accounting software or searching for apartment listings, show the potential for AI to handle routine but multi-step online tasks. However, the feature's availability is currently limited to Plus, Pro, and Business plans, which may restrict access for casual users.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
On August 25, 2026, OpenAI unveiled Jalapeño, its first custom inference chip, claiming 1.5–1.9x more AI work…
The Linux Foundation announced the contribution of TRACE (Trust, Runtime Attestation and Compliance Evidence)…
A new Stanford research paper warns that AI chatbots and agents may let advertising dollars influence what the…
The Wall Street Journal published an op-ed by billionaire investor Stanley Druckenmiller that was written with…
OpenAI unveiled benchmarks for its Jalapeño inference chip on Tuesday, showing it competes with Nvidia’s best…
Researchers from Bytedance Seed developed EdgeBench, a benchmark measuring how AIs improve on tasks over multi…
