The Goldilocks Time Window for Proactive Interventions in Wearable AI Systems

TL;DR

The paper tackles the challenge of delivering proactive interventions through wearable AI without causing intrusiveness or dependency. It introduces the Goldilocks Time Window, a contextually adaptive timing concept, and identifies factors such as social sensitivity, modality, and goal alignment that shape optimal intervention moments. Through prototypes and case studies, the work demonstrates how timing mechanisms like a debouncer can balance responsiveness with user autonomy, and it calls for evaluation frameworks that prioritize timing efficacy and ecological validity. The expected impact is toward more effective, user-aligned wearable AI that intervenes precisely when beneficial, improving real-world outcomes while maintaining trust and agency.

Abstract

As AI systems become increasingly integrated into our daily lives and into wearable form factors, there's a fundamental tension between their potential to proactively assist us and the risk of creating intrusive, dependency-forming experiences. This work proposes the concept of a Goldilocks Time Window -- a contextually adaptive time window for proactive AI systems to deliver effective interventions. We discuss the critical factors that determine the time window, and the need of a framework for designing and evaluating proactive AI systems that can navigate this tension successfully.

Figures (4)

• Figure 1: The Goldilocks Time Window
• Figure 2: Overview of Mirai.
• Figure 3: Memoro's interaction modes: (1) Query Mode where the user can ask contextual questions (2) Queryless Mode where the user can request predictive assistance and skip query formation. In both modes, responses are discreetly played back to the user using a bone conduction headset.
• Figure 4: An assistive bone-conduction headset for people with visual impairment.