On the 28th of October, @manetta and I will be giving a at the Impakt festival. Show more
During the trust <in> formation workshop we will take the ORES project as an inspiration for machine learning done ethically: transparency in the decision making process and stimulation of multiplicity of uses.
Instead of "post-truth" algorithms, we'll make the jump to "post-trust" algorithms: the workshop looks at the reasons why Wikipedia is currently seen as a trustable source of information mainly built through collaboration. It's not only trusted by humans, but also by machine learning developers: many models use its articles to train or test on. The fact that Wikipedia has its in-house team working on machine learning is a less known fact: the ORES project.
The workshop will provide a practical experiment in which the participants are invited to write together under the watch of the ORES models and prove their good faith.