Verified, then logged.
Sun 22 Nov 2026 · 00:19 KST / 15:19 UTC한국어 | ENSearch /

02 / IT · Analysis

ETRI develops Trinity to assess user-aware task planning

Trinity assesses whether a planned action suits the person receiving the service, beyond whether it completes the instruction, AI Times reported.

An analyst seen from behind compares two side-by-side monitors showing the same video stills with objects outlined differently on each screen in a dim office. AI-generated image

ETRI, South Korea's Electronics and Telecommunications Research Institute, developed Trinity to assess task planning that considers people and their surroundings, AI Times reported. The report described an evaluation benchmark for robots and AI agents that examines whether their choices suit the user.

AI Times used dietary habits and allergies to illustrate the distinction between following an instruction and choosing an appropriate action. Those examples put the person receiving the service at the centre of the assessment.

The report raised a question about what counts as successful performance: whether completing an instruction accurately is sufficient when the user's characteristics also matter. That shifts the discussion from completing a task to choosing how to carry it out.

Task planning that considers people and surroundings

According to AI Times, ETRI announced Trinity as a way to evaluate the task plans of robots and AI agents. The assessment considers both personal characteristics and the surrounding environment.

The report described Trinity as a benchmark, a standard used to assess capabilities. Its focus brings the context of a task into the evaluation alongside the ability to perform it.

Appropriate actions for the user

AI Times distinguished task competence from the suitability of an action for the person receiving the service. Its dietary and allergy examples showed why an assessment also needs to consider whom the action serves.

The report's central question concerns actions that complete a task while overlooking the user's characteristics. The resulting standard broadens the assessment from whether a task is finished to whether the chosen action fits the person and the setting.