Be Happy to Stop the Alarm

ESAD Orleans, 2025-06-12

Project DNA (diplôme national d’art) with honors (Félicitations)

We all struggle to get up in the morning, and a conventional alarm clock does not necessarily provide a pleasant wake-up experience. Based on this observation, I created Be Happy to Stop the Alarm, an alarm that can only be turned off by smiling. The smile is detected using artificial intelligence.

To design it, I disassembled a real mechanical alarm clock in order to integrate a Raspberry Pi and a mini camera. For facial recognition, I used TensorFlow Lite as the execution engine and Python for the script. I personally trained a model capable of detecting facial emotions using 28,709 training images. To do this, I followed a tutorial and used the dataset provided by the YouTuber Digital Sreeni.

I am inspired by the idea of human–machine co-evolution described by Bruce Mazlish in The Fourth Discontinuity. By adding humor, I transform ordinary daily tasks into unfamiliar experiences, drawing from Kenji Kawakami’s concept of Chindogu—objects that propose seemingly perfect yet absurd solutions to everyday problems.

Through this process, I realized that humor in my work also carries an ironic critique of AI’s growing presence, balancing positivity with a potential loss of autonomy in society. This perspective echoes Modern Times by Charlie Chaplin and relates to Valia Fetisov’s Installation of Experience. Together, these ideas suggest moments of constraint, subtle coercion, and even a quiet sense of dystopia.