The Montreal Declaration for a Responsible Development of Artificial Intelligence was produced by the Universite de Montreal and released in 2018. Its co-construction process had begun on 3 November 2017, and the finished declaration was published about a year later after extensive public consultation. The effort was led by two scientific co-directors, philosophy professor Marc-Antoine Dilhac and design professor Christophe Abrassart, and it became an early and widely cited example of bottom-up, participatory AI ethics.
The declaration is built around ten principles, drawn from a set of underlying values: well-being, respect for autonomy, protection of privacy and intimacy, solidarity, democratic participation, equity, diversity inclusion, prudence, responsibility, and sustainable development. Alongside the principles, the document offers 59 detailed recommendations meant to translate the values into practice for developers, companies, and governments.
What set the Montreal Declaration apart was its method. Rather than being drafted behind closed doors, it emerged from an inclusive deliberation process involving more than 500 citizens, experts, public officials, and industry and civil-society stakeholders across a series of workshops and public forums. That open process made the declaration a reference point for later efforts to write AI ethics with public input rather than only expert input.