IGF 2018 WS #198
Accountable and trustworthy algorithms

Issue(s)

Organizer 1: Simon Chignard, Etalab - Dinsic
Organizer 2: Véronique Routin, Public innovation / French national school of administration (education & research)

Speaker 1: Jaimie Boyd, Government, Western European and Others Group (WEOG)
Speaker 2: Simon Chignard, Government, Western European and Others Group (WEOG)
Speaker 3: Véronique Routin, Civil Society, Western European and Others Group (WEOG)
Speaker 4: Guillaume Chaslot, Technical Community, Western European and Others Group (WEOG)

Moderator

Véronique Routin

Online Moderator

Véronique Routin

Rapporteur

Simon Chignard

Format

Panel - 60 Min

Interventions

Guillaume Chaslot (Algotransparency / Center for Human Technology) is a former Youtube engineer. He had questioned the mechanism of Youtube recommandations systems. His work has been featured on The Guardian, Wired.
In this session, he will provide an insight of the role civil society can play in monitoring the use of algorithm.

Jaimie Boyd is in charge of the Canadian open government policy. She is an advocate of algorithmic accountability inside the federal government of Canada. In this session, she will explain how this topic articulates with the "Open Canada Policy".

Simon Chignard (Etalab / Dinsic) is senior strategic advisor at the French Prime minister task force on open data, open source and algorithms. In this session, he will share his experience on designing transparent, ethical and accountable algorithms, and specifically of Parcoursup, the allocation platform used by 900 000 students every year that he worked on.

Diversity

Diversity in this session is promoted by a gender balanced panel, the participation of government as well as civil society and academics representatives.
Speakers come from two continents (Western Europe and North America).

A short introduction (10 min) by the moderator Veronique Routin will introduce the subject of accountability in algorithmic decisions and outlines the key questions of this session: why should we hold algorithms accountable ? How to make sure that the decisions (or recommandations) of the algorithm is understandable by the user and, more generally, can be debated by the society ? Is accountability the key to reinforce trust in these systems ?

Each speaker will respond to the questions by providing the attendance with some insights based on their own experience in diverse situations and contexts.

After the first round of insights from the speakers, the attendants will be given the ability to ask questions through a circulating microphone. The moderator may or may not try to specify the question so as the speakers can give the most interesting answer to it.

A growing number of algorithms are taking decisions or making pivotal recommandations. Some take mandatory decision, some are just using the power of suggestion and incentives. Some are private, other public. But the key question today is how to make sure they can be held accountable ?

As the use of algorithms may have a tremendous economic, social and societal impacts on the citizens, addressing their governance further asks the question of the future of the Internet. Speakers will especially have raise the question of knowing if there's still a place for opaque or non-accountable algorithms within modern and connected democracies.

Online Participation

The online participation will be encouraged through a live broadcasting. The online attendants will be able to ask live questions that the moderator could pick to raise the evoked issues.