Cambridge, MA, U.S.
An Agentic AI Radio for Ambient, Physical, and Communal Interactions
Core77 Design Awards 2026
Emerging Technologies Award
Student Runner Up
3.6.2026 - 3.8.2026
(48 Hours)
Hardmode Hackathon
MIT Media Lab
Collaborator :
Lilith Yu, Chloe Ni
Role:
Concept Development,
Hardware,
Project Management
# Agentic AI
# Physical AI
# Human-AI Interaction
# Agent2Agent
# Product Design
# Interactive Design
Project Description
What if the best way for AI agents to find each other wasn't a software protocol, but a stranger handing you a key?
People used to feel connected just by tuning into the same thing at the same time. For most of the twentieth century, the household radio was exactly this: ambient, always on, belonging to the home rather than to any individual. It connected people across distance not through direct messaging but through the act of tuning in and discovering each other.
Radio Ga Ga revives this quality for the age of agentic AI. It is an AI-powered household radio that embeds a conversational agent inside a vintage radio form factor, using analog controls and physical tokens to deliver live, context-aware audio content while quietly establishing connections between people.
The project proposes an alternative to the dominant direction of AI hardware. It asks what agentic AI might become if it belonged to a home instead of an individual. Instead of making intelligence more individualized, optimized, and attention-hungry, it reframes AI interaction as ambient, shared, and socially situated.
Context
Today’s AI devices are largely imagined as personal companions: in your pocket, in your ear, or attached to your body. We were interested in the opposite model. Historically, radio was ambient, shared, and always available without demanding focused attention. It connected people not by direct messaging, but by letting them tune into the same stream at the same time. Radio GaGa brings that social and atmospheric quality into the age of agentic AI.
We believe the most compelling form for embodied AI is not the personal device that isolates the user, but the ambient object that blends into shared life.
Radio GaGa uses a familiar household form to make agentic systems more understandable, less demanding, and more socially situated.
Interaction Design
The interaction model is deliberately physical and low-friction. There is no touchscreen, no app, no wake word, and no companion interface. Users approach the radio, turn the dial, and listen.
A central control maps what we call a radius of attention: a spectrum from private to shared to public. Rather than opening apps or issuing commands, users navigate a continuous field of AI-generated content through familiar radio gestures.
An e-ink display provides just enough information — channel, topic, and time — to orient the listener without pulling the device into the attentional category of a screen.
Connection Mechanism
Connection between radios happens through NFC household tokens. Instead of making social linking instant and weightless, we wanted connection to feel tangible and intentional. One household hands a token to another; once tapped, their radios begin an ongoing agent-to-agent relationship.
This gesture is less like sending a follow request and more like exchanging house keys. It turns inter-agent communication into a human social ritual.
System Logic
Behind the interface, Radio GaGa runs on a Raspberry Pi 5 with an asynchronous architecture that coordinates hardware input, LLM streaming, and real-time audio playback. Distinct host personas generate contextual radio segments using live data such as weather, news, sports, and events, then stream them into voice output.
To preserve the feeling of live radio, content is pre-generated in the background so that switching channels feels immediate. Peer-to-peer communication between radios allows households to maintain lightweight inter-agent relationships over time.
Radio GaGa
@MIT Hardmode Hackathon
Radio once gave us a shared pulse. Now, our AI agents are silos.
We’re breaking the isolation with an AI radio: multi-agent hosts for you and your friends.
Tune in together.