Teleoperation
A person controlling a machine remotely, which is a legitimate design and is also what partially autonomous systems look like when nobody publishes the intervention rate.
When not to use it
- As a criticism. Teleoperation is often the right answer for rare, high-stakes or unstructured tasks.
- Where latency makes the control loop unstable, in which case local autonomy is required regardless of preference.
- As a permanent substitute for autonomy in high-volume tasks, where the labour cost does not amortise.
Reach for something else instead
- Full autonomy — appropriate where the task is bounded and the environment controlled enough to make it reliable.
- Fixed automation — where the task never varies, removing the need for either autonomy or an operator.
- -
Read more on the blog
- 2.6 million robotic surgeries, none of them autonomousThe most-deployed medical robot makes no decisions. And the largest meta-analysis of its outcomes was co-authored by the company that sells it.
- 32 million vacuums, and 38% on household tasksThe most-deployed consumer robot succeeded by accepting a worse result more often. The home is the one place where the other ways of making a task tractable are unavailable.
Further reading
- International Federation of Robotics, World Robotics 2025 — the deployment context in which autonomy claims are made.
- Parasuraman & Riley (1997), Humans and Automation — the supervisory role and its failure modes.
Primary sources, listed so you can check the claims on this page rather than take them on trust.
Where people go wrong
- Reading a demonstration as evidence of autonomy when no intervention rate was reported.
- Comparing systems on task success without controlling for how much human assistance each received.
- Assuming shared autonomy is a transitional stage rather than, frequently, the correct final design.
At a glance
Where this sits
A starting point. Nothing needs to come before it.
Computed from the prerequisite graph, not assigned. How this works