Back to AI News
arxiv

Agents Can Read the Clock but Still Prep Users for Expired Events, Study Finds

Researchers frame 'temporal state perception and use' as the ability of long-horizon agents to perceive the current time and apply it when interpreting dated records left over from earlier conversations. In session-rekindling tasks spanning five model routes and five temporal-information conditions, agents reliably queried the clock and even noted that a scheduled date had passed, yet many still told users to prepare for the expired event. Supplying explicit temporal context reduced this error, and providing the computed date relation removed it entirely. The authors release a session-rekindling benchmark and code to inform temporal-awareness guidance for long-horizon agent design.

AI Brief

It pinpoints a concrete, measurable failure mode in persistent assistants — perceiving elapsed time without acting on it — and delivers both a benchmark and actionable mitigation guidance for agent builders.

arXiv:2611.17806v1 Announce Type: new Abstract: Long-horizon agents increasingly serve as persistent personal assistants across modern web platforms, managing schedules, applications, and multi-step plans over weeks or months. Users often return to an ongoing conversation after a long conversational hiatus, while earlier bookings and preparation notes stay on record. During this hiatus, scheduled dates may pass even though the retained records remain unchanged and event outcomes go unrecorded. Existing evaluations of long-horizon agents focus on fact retention and explicit state updates. Whether agents perceive the passage of time and adapt their advice accordingly therefore remains unknown. We frame this ability as temporal state perception and use, which means perceiving the current time and applying it when interpreting dated records. To study it, we design session-rekindling tasks that keep the earlier conversation fixed and vary only the follow-up date and the way temporal information reaches the agent. Experiments across five model routes and five temporal-information conditions show that agents can obtain the current time but often fail to use it. They reliably query the clock and even note in their reasoning that the scheduled date has passed, yet many still tell users to prepare for the expired event. Supplying explicit temporal context reduces this error, and providing the computed date relation removes it entirely. Keeping up with time therefore depends on turning perceived time into updated plans. Based on these findings, we design a session-rekindling benchmark that provides temporal-awareness guidance for designing long-horizon agents. The code can be found at https://github.com/Adoreu2/session-rekindling-benchmark.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!