The safe algorithm
The Safe Algorithm challenges the gender binary inherent in facial recognition systems by proposing a “safe digital training space” that reimagines how artificial intelligence can represent human diversity.
By training the model with images from the queer community, an inclusive database is created, which then feeds back into the system to refine its perspective, producing broader representations of gender.
The result is a virtual environment that engages with the concept of a “safe space,” integrating both queer and heteronormative individuals as a tool for critical reflection. In doing so, it not only challenges the technical foundations of algorithms but also redefines the notion of safe spaces as bridges for connection, fostering dialogue between communities and rethinking the machine-human relationship.
