Can workers tell apart a task that feels pointless from one that's simply confusing — and does that distinction show up in their ratings?

This explores whether people can tell a task that feels pointless apart from one that's just unclear, and whether that difference shows up in the preferences they report. No study in the collection puts both kinds of task in front of workers and asks them to sort them. Two separate threads come close. One measures meaninglessness directly. The other gives a way to spot ambiguity hiding inside people's ratings.

On the meaningless side, researchers built a five-item 'bullshitness' scale and used it with 202 workers rating 171 tasks across 22 occupations Does perceived task meaninglessness drive worker desire for AI delegation?. The scale defines meaninglessness by two things: the task feels pointless, and it seems disconnected from what the organization is trying to achieve. It does not define it by the task being hard to understand. That matters because a meaningless task can be perfectly clear. People know exactly what to do and just don't see why. Those ratings predicted behavior. The more pointless a task felt, the more workers wanted to hand it to AI and the less they wanted to oversee the result. They also wanted something specific from the agent: speed, simplicity and directness, not explanations or creativity What agent traits do workers prefer for meaningless tasks?. In short, workers want meaningless work finished and gone.

Ambiguity looks different, and the best lens for it comes from research on preference annotation, not workplace surveys. That work splits people's responses into three kinds Do all annotation responses measure the same underlying thing?. Genuine preferences stay stable. Non-attitudes are answers given when the person has no real view. Constructed preferences are made up on the spot and shift with how the question is framed. You can tell them apart by checking whether the same answer comes back when you ask in a different way. An ambiguous task tends to produce non-attitudes or constructed preferences, because people are improvising. A meaningless task can still produce a firm, consistent answer: 'just get it done.' So the useful signal isn't the rating itself but whether it holds steady.

One caution: what people say they feel doesn't always match what they understand. In conversation studies, users reported being satisfied while they were still confused, especially when they didn't know what they were missing Does user satisfaction actually measure cognitive understanding?. A worker who rates an unclear task as pointless may really be reacting to confusion, and nothing in a single rating would show it.

Put together, these suggest a test the corpus implies but never runs. If a worker's 'this is pointless' rating stays the same however you ask, the task is probably meaningless. If it changes with the wording, the task is probably ambiguous. This matters for AI delegation. Workers are happy to hand meaningless tasks to a fast, low-oversight agent. An ambiguous task mistaken for a meaningless one would get exactly that agent, with no one checking the work.