AI for everyday life — öiō × CIID

01

Context

This project was made during a CIID workshop with Matteo Loglio and Bjørn Karmann from öiō studio.

The focus was working hands-on with AI & ML to explore Ambient Intelligence.

The main concept was: Thinking homes.

AI and ML can bring objects to life, making them react and adapt to different stimuli from both the users and the environment.

An open balcony door in the villa, framing Lake Como and the mountains beyond.

A projected slide: “Smart homes” struck through, an arrow pointing down to “Thinking homes”.

The empty frescoed hall used as the workshop studio, with tall windows onto the garden.

The group at café tables on the lakeside terrace during a break.

the “office”

02

Goal

Our first task was to look around the place, find some objects, give them a name, write some background stories about them.

Then, my teammate and I chose one object, and got assigned a random story.

Finally, our goal was to bring it to life with AI tools, and any material we could get our hands on.

The wrought-iron gate of Villa del Grumello in Como, with the villa's name on a sign.

A workshop participant leaning against a pine trunk with a small green lantern fixed to it.

Two palm trees standing side by side in the villa garden, the lake behind them.

A rock outcrop overgrown with ivy, its ledges shaped roughly like a seat.

some of the “objects”

03

Outcome

My teammate and I ended up bringing to life a statue of Ugo Foscolo — with PTSD.

Ugo Foscolo, Italian poet

A road was nearby, and many small planes were taking off and landing on the lake; those noises became the PTSD triggers.

To visually express fear, we added one animated arm using an Arduino UNO R4 and a servo motor (leaving the statue untouched). Since he was a poet, we also gave him a voice through a bluetooth speaker.

When noise levels were high, the statue covered its face and recited poems with a custom voice. The performance ended, with the help of ML and a webcam, when someone hugged the statue.

The marble bust of Ugo Foscolo in the garden, a pale cardboard arm raised across its face.

what it was built with

An Arduino UNO R4 WiFi wired to a 25 kg servo, a copper pipe-cleaner arm attached to the horn.

Google Teachable Machine: sample classes on the left, training in the middle, a live preview on the right.

The ml5.js fast style transfer example, an input photo restyled by a model in the browser.

The p5.js web editor: a short sketch on the left, its generated drawing on the right.

The ElevenLabs workspace: recorded takes above, voice waveforms below, a row of language flags.

The Cohere playground, an empty chat waiting for a prompt.

04

The interaction loop

The statue behaves like a small state machine, sensing through a laptop's mic and webcam, with ML running locally in the browser (Teachable Machine + p5.js):

Listen — ambient noise is continuously classified on-device
Trigger — loud noise (planes, traffic) wakes the statue
Perform — a servo-driven arm covers its face (Arduino UNO R4, over WiFi) while it recites poems through a BT speaker
Release — when the webcam recognizes a hug, the performance ends

The final demo at golden hour: the laptop driving the piece on a folding chair, the bust lit from behind.

full video

05

Conclusions

AI can turn seemingly intricate interactions with objects into natural experiences — the interface disappears into the environment.
On-device ML is far more accessible than expected: training and running a model no longer requires a research team.
AI tools amplify a designer's reach, but only with a clear intent: they execute, they don't decide.
Designing interactions that live in physical space — timing, sound, movement — is a different craft than screen design, and I want to do more of it.

An Instagram story by öiō studio captioned “A statue afraid of loud noises”, the arm raised over the bust.

The two of us either side of the bust at the end of the workshop, its cardboard arms spread wide.

öiō × CIID | AI for everyday life