The core constraint facing phone-based AI assistants is context continuity: a model that shuts off when the screen locks is not ambient. Glasses, worn continuously and positioned at the human face, keep microphones and cameras live in a way that a pocketed phone cannot replicate. Big Tech has made that property the centerpiece of its argument that wearables are the gateway to ambient AI, and critics have answered with a specific charge: cringe stalkerware.
The gateway claim rests on a specific technical premise. Continuous access to audio and camera input, from a device at eye level, gives AI a sensor window that no handheld device provides by default. Big Tech's framing positions glasses as the persistent AI interface layer, a role the phone cannot fill by design.
The criticism accepts the technical logic. The target is the data architecture that logic requires. Capture that runs continuously at eye level means bystanders, conversations, and physical spaces are logged without their consent. The "cringe" modifier in "cringe stalkerware" bundles two objections: social awkwardness and the surveillance structure embedded in always-on capture. Earlier wearable computing drew the first. This generation has earned the second.
The standoff traces to a single feature. Persistent ambient sensing is both the product's value proposition and its primary privacy exposure. Those two properties are not separable at the hardware level, which means resolving the dispute depends on policy and governance decisions that neither Big Tech nor its critics has addressed in public.