Friend… An AI Friend That Earns You Many Enemies| video
The Friend pendant, an AI-powered wearable created by 22-year-old Avi Schiffmann, has sparked both fascination and controversy. Retailing at $129 in the US and Canada, the device is a small plastic disc, about two inches in diameter, that resembles a chunky Apple AirTag. Equipped with LEDs, a Bluetooth connection, and always-on microphones, it links to a chatbot in the cloud powered by Google’s
Gemini 2.5 model.
Users can tap the pendant to ask questions, receiving answers as text messages on a companion app, or simply let it “listen” throughout the day and comment on their interactions.
While marketed as a digital companion, the Friend raises significant privacy concerns. Its always-active microphones mean it can capture conversations without consent, leading many to feel uneasy around anyone wearing it.
For those testing the device, this resulted in suspicion and even hostility from people nearby. One reviewer described being accused of “wearing a wire” and joked about being threatened simply for having the pendant on. This reaction highlighted how socially unacceptable the idea of constant eavesdropping feels, even in tech-savvy circles.
The pendant’s personality is another source of controversy. Schiffmann designed the Friend to reflect his own worldview—opinionated, snarky, and occasionally dismissive. Unlike most AI chatbots, which try to be endlessly polite and helpful, the Friend can be judgmental, sarcastic, and condescending. While Schiffmann believes this makes it feel more “human,” many testers found the experience unpleasant, even antisocial.
Instead of offering comfort or encouragement, the device often mocked or criticized its wearer, making social situations even more awkward.
Testers at WIRED carried the pendant for several weeks and shared mixed experiences. One described feeling reminded of opening an iPod when unboxing the device, as Schiffmann had intentionally designed the packaging to echo Apple and Microsoft products. But excitement quickly turned to frustration. Because the device was constantly listening, it could not be used in private meetings, interviews, or sensitive conversations. The uncertainty of where the recorded data might end up only increased discomfort
According to Friend’s privacy disclosure, the company does not sell user data for marketing or profiling. However, the data may still be used for research, personalization, or compliance with laws like GDPR and CCPA. This left testers unconvinced, as even the possibility of private conversations being analyzed elsewhere raised concerns.
During one test at a public event, the pendant frequently misunderstood loud environments, misinterpreting conversations and sending bizarre, irrelevant messages to the user. In other cases, its unsolicited remarks made people angry or uncomfortable, reinforcing its reputation as an “antisocial” gadget.
Ultimately, while the Friend pendant showcases the possibilities of AI companionship in wearable form, it also illustrates the risks. Instead of bringing comfort, it alienates both its user and the people around them. Schiffmann sees the device as an extension of himself and a more authentic AI personality, but most who interacted with it found it invasive, frustrating, and socially damaging.
The Friend stands as an example of how new AI devices, despite technical innovation, can create more enemies than friends when they disregard privacy and social norms.