KlueIQ is built on human–AI collaboration, but the standard language for that relationship is terrible.
“Robot” comes from robota: forced labour. “Automation” sounds like a headcount reduction memo. Even “tool” quietly erases whose labour is being used to train and run these systems.
So KlueIQ uses three terms on purpose:
- Agent: any non-human system acting in the world: an AI model, a scoring engine, a recommendation system, a scripted process. “Agent” is structural; it says what something does, not what it’s worth.
- AI collaborator: a specific agent you are in an active relationship with on a task: co-writing, co-analysing, co-designing. This is the word we use when the work is genuinely shared.
- Aite: our house name for a human–AI creative pair on a specific project: you + your AI collaborator(s) acting as one composite maker.
An aite is not a robot and not a replacement human. It is a configuration: a person, plus at least one agent, plus a shared goal. The “AI” in aite is doing pattern detection, retrieval, synthesis, simulation. The “human” is doing judgement, taste, moral reasoning, and final say. The aite is the unit we care about when we design experiences.
What this vocabulary commits us to:
- We will not talk about “robots” doing people’s jobs. We will talk about agents embedded in systems that still have human responsibility attached.
- We will talk about collaborators when there is real co-creation, not when an API call fills in a spreadsheet.
- We will treat every aite as accountable to the human in the loop. If something goes wrong, responsibility routes back to the person and the institution, not the agent.
Language does not solve the power issues around AI, labour, or surveillance. But it does reveal which side you think you’re on. KlueIQ is explicitly on the side of humans building with agents, not being graded by them.