When AI ruled France – Macron’s large scale consultation

In 2009, as a response to the “mouvement des gilets jaunes” (yellow vest protests) in French, President Macron instigated a “Great National Debate”, a series of face-to-face meetings across the country accompanied by a web-based systems to solicit widespread views.  This exercise gathered more than 300,000 responses, far too many to deal with manually, so a company specialising in text analysis was recruited to extract key themes and issues [bibtem name=”Bo19,Fr19″].  These were then used, as a form of focus group politics in the large, to shape French Government policy.  One could say France was ruled by AI!

Manifestation du mouvement des gilets jaunes de la FFMC, à la place de la République, à Belfort, le 17 novembre 2018.

It is worth noting here that despite Macron’s efforts to gather views – there was no accountability put in place – e.g. views were gathered but then no mechanisms for people to check how any of their views translated in policies.

Imagine how this could have been different, maybe the ability to view your submission and see how it contributed to various high-level themes.  This would be possible even with complex machine learning algorithms (a rarer form of explainable AI), and fairly straightforward with most forms of text analysis.

References

[bibtem name=Bo19] Anne Bouverot (2019). Great National Debate: Artificial Intelligence at the Service of Collective Intelligence?
Interview with Anne Bouverot. Expressions by Montaigne 12/04/2019.  Institut Montaigne. https://www.institutmontaigne.org/en/expressions/great-national-debate-artificial-intelligence-service-collective-intelligence

[bibtem name=Fr19] France 24 (2019).  What will France do with ‘National Debate’ data?  France 24.  03/03/2019. https://www.france24.com/en/20190302-france-great-national-debate-data-artificial-intelligence-politics-yellow-vests