Companion robot privacy comparison
Privacy cannot be reduced to one score. Start with sensors, then ask where processing happens, what requires an account and what continues to work if cloud services disappear.
The simplest privacy profile: no camera, no cloud
Qoobo, Petit Qoobo and fufuly avoid the largest sensor/cloud trade-offs because their core interaction is tactile and local. Moflin is also much less camera-centric than a mobile AI robot. This simplicity comes at the cost of open-ended conversation and remote features.
Ropet publishes strong local-processing claims
Ropet says everyday interaction can run offline and normal image/audio processing can occur on-device. Optional cloud features can still change the data path, so privacy should be evaluated feature by feature.
Loona, Miko, EMO/AIBI and aibo are sensor-rich by design
These products gain capability from cameras, accounts and connected intelligence. Sony aibo in particular is explicitly tied to an AI Cloud Plan for its full adaptive experience. That does not make them inherently unsafe; it means the current privacy policy and service architecture are material parts of the purchase.
Enabot is a different privacy problem
EBO and ROLA products are primarily mobile camera/remote-presence devices. Their value depends on video, audio, remote access and in some cases cloud or AI services, so buyers should treat account security, storage, app permissions and remote-access controls as first-class buying criteria.
Our pre-purchase checklist
Check camera and microphone presence, local versus cloud processing, account requirements, retention and deletion controls, child-data rules where relevant, remote-access security, optional cloud plans, and what functionality remains if the vendor service is unavailable.