Apple explores user control over personal sensing systems in new HCI paper

Apple’s Machine Learning Research group published a paper titled 'Negotiating Ontological Boundaries in User-Authored Personal Sensing Systems,' which examines how designed artifacts shape and limit what users can achieve with personal sensing systems. The authors argue that current evaluations of such systems often focus on usability, usefulness, or technical feasibility, while neglecting how these systems influence what users can imagine or accomplish.

The researchers designed two open-ended probes that employed a Wizard of Oz technique to simulate the experience of training a personalized machine learning system on phenomena defined by participants themselves. This approach allowed users to explore and define the boundaries of their own sensing systems in real-world contexts.

In a week-long exploratory study, participants used one of the two probes during their everyday lives. The study identified four key sites where ontological boundaries were negotiated: the boundaries of a phenomenon, the role of the subject in relations, the distinction between signal and noise, and the objectivity of data.

The authors propose starting points for supporting boundary negotiation through design and discuss the use of open-ended probes as a method for ontological design. The paper includes contributions from researchers affiliated with Apple and Stanford University, including Nava Haghighi, Danielle Olson, Halden Lin, Erdrin Azemi, Gierad Laput, Kayur Patel, and James Landay.

The research was published in October 2026 and is available on Apple’s Machine Learning Research website. The paper also includes related readings and updates, such as a 2026 workshop paper on behavioral privacy leakage in agentic negotiation and a 2025 system description for ODKE+, an ontology-guided knowledge extraction system.