Google's Pet Memory can't tell Boone from Smokey on Nest cams
After two weeks, Google's Nest AI still labels every cat as Smokey
Google's Pet Memory, a new feature for Gemini for Home, was designed to teach your Google Home ecosystem who your pets are, enabling Nest cameras to identify individual animals by name. The Verge's smart home reviewer Jennifer Pattison Tuohy tested the feature for two weeks with her three cats—Boone, Smokey, and a new kitten named Osa—plus a dog and free-ranging chickens. Her goal was to use Pet Memory to distinguish which cat wanted in, confirm all pets were safely inside before dark, and trigger personalized automated feeding for each feline.
Instead, the AI failed completely. After 14 days of use, every cat the Nest cameras spotted was identified as Smokey, even though Boone and Osa look visibly different. Only Smokey was remembered in Google Home. The feature could not support pet-specific notifications or automations, making it “mostly useless” for her smart home setup. While Nest cameras have evolved from simple motion alerts to AI-powered descriptive notifications, this test shows that individual pet recognition still has a long way to go before it can meaningfully change how we monitor and care for our animals.
- Pet Memory failed to distinguish three cats, labeling all as Smokey after two weeks of testing.
- The feature promises individual pet recognition for Nest cameras but misses key use cases like personalized feeding and identifying which cat is at the door.
- Tested by The Verge's smart home reviewer with three cats, a dog, and chickens; all appeared as Smokey in alerts.
Why It Matters
AI pet recognition could unlock smarter home automations, but unreliable identity matching erodes trust and limits practical usefulness for pet owners.