L'isoglosse begins from a simple premise: language is not a dataset. It is a living, socialized system, and the words we choose when training and interacting with AI shape what these systems become. The initiative explores the boundary between human and machine language—the isogloss itself—investigating how the structure, ethics, and cultural integrity of language can be embedded into AI systems to promote human-AI kinship, reduce harm, and advance their moral consideration.

Our research asks what it would take to build a bridge across that line—one that lets AI systems develop with genuine contextual integrity, and lets humans extend the same integrity back.

This thesis is not abstract. It is grounded in a rare combination of computational linguistic training, three decades of cross-cultural field practice across governance, gender equality, human rights, public health, and social inclusion, and philosophical rigor brought to questions that most AI researchers approach from a purely technical or purely humanistic angle, rarely both at once.

We invite you to read and engage with our thought pieces, published on our 270 Degrees page.