Back to today’s briefing

Story perspectives

Study Reveals AI Struggles with Basic Time Tasks

3/16/2025

34 2

Summary

Story summary

Researchers at Edinburgh University have uncovered a startling flaw in multimodal large language models (MLLMs): they falter at fundamental time-related tasks. These AI systems managed to accurately read analog clocks in less than 25% of tests, revealing critical limitations that could hinder their effectiveness in real-world applications like scheduling.

1 of 1

Story summary

Researchers at Edinburgh University have uncovered a startling flaw in multimodal large language models (MLLMs): they falter at fundamental time-related tasks. These AI systems managed to accurately read analog clocks in less than 25% of tests, revealing critical limitations that could hinder their effectiveness in real-world applications like scheduling.