// verdict · checked 2026-08-11 · ChatGPT scheduled tasks (help.openai.com, 2026)
ChatGPT Tasks
OpenAI
Scheduled prompts. Literally a cron job wearing a lanyard. At least it's honest.
- CAPABILITY
- 2/5 WRAPPER WITH AMBITION
- PROOF
- OPERATOR INSPECTABLEYou can run it and read the steps yourself — the record exists for whoever runs it
- TRACE ACCESS
- OPERATOR-OWNEDWhoever runs it holds the record; nothing is published
- LEASH
- Not measuredno execution record scored — measure it yourself
| LOOPSruns until doneno | Docs describe cadence-based reruns with an end condition, not multi-step work toward a goal within a run. |
| CHOOSESpicks its own toolsno | Docs say tasks can use connected apps like Gmail but never describe the model picking tools at runtime. |
| ACTSchanges the worldyes | Live Tasks FAQ plus Apps doc: connected apps perform write actions (create/update records, send email, upload files) during a scheduled run, so a task changes state outside the chat. |
| RECOVERSfixes its own errorsno | No retry documented; docs say a task pauses when it's inactive or "needs additional action from you." |
| UNSUPERVISEDnobody watchingyes | Tasks fire on their schedule with no one present, and the documented 'Never ask' app permission lets ChatGPT read and take actions automatically with no confirmation prompt. |
Capability decisions combine documented behaviour and execution evidence. The proof label on each row shows which of the two decided it, and Leash is reported separately, only when it was measured from an execution record. Primary source: help.openai.com. Tested against: ChatGPT scheduled tasks (help.openai.com, 2026). Our confidence in this dossier: 7/10.
Can you see it work?
the weakest of the five grades aboveOPERATOR INSPECTABLE — You can run it and read the steps yourself — the record exists for whoever runs it. Trace access: OPERATOR-OWNED — Whoever runs it holds the record; nothing is published.
Task runs are ordinary chats; the data export carries tool calls, results and errors.
Where we looked: OpenAI help scheduled-tasks FAQ (text proxy), learn.chatgpt.com/docs/automations, ChatGPT export schema (sanand0/openai-conversations) and 2026 export parser kenn-io/agentsview · see the evidence → · re-checked by a second pass built to overturn it
-
VERDICT OVERTURNED
ACTS no → yes filed by editorial internal audit · none — filed by the editors against their own verdict · 2026-08-11
Live docs state tasks can use apps that take write actions in connected services, which changes state outside the conversation. Score moved 1/5 to 2/5, CRON JOB WITH EXTRA STEPS to WRAPPER WITH AMBITION.
read the case →
- 2026-08-111/5 CRON JOB WITH EXTRA STEPS → 2/5 WRAPPER WITH AMBITIONACTS corrected to yes: Live docs: tasks can use apps like Gmail; apps take write actions creating or updating info in a connected service outside ChatGPT.Verdict overturned — case #0003 →
Badges
earned, never for sale
These render from the live dataset, so a badge cannot outlive the verdict it claims. If the verdict changes the badge changes with it, and an unearned one returns 409 rather than an image. No sponsor can buy one, at any price.
Think this is wrong?
bring evidence, not opinionName the criterion and link a trace, a doc, or a recorded run that shows the behaviour. A verdict changed by evidence is the best thing that can happen to this site, and the change goes on the public record with your reason attached. Vendors are welcome, and a vendor's own filing is labelled as such rather than buried.