ARTICLE | AI and politics: In France, startups are indicating the market to follow

Article published in French in Libération.

590 French start-ups specialized in artificial intelligence (AI), including 76
that deal with generative AI – which creates new content – and not a single
one directly linked to politics. In its latest census, the digital industry
association France Digitale, a lobby for digital companies, identified
around twenty different sectors. But does this mean that there is no
market for AI in politics? “No, there is a market, but companies don’t
advertise politics as their primary customers. That’s not how they define
themselves,” observes Marianne Tordeux and Thomas Barreau at France
Digitale.

“Generative AI will disrupt the world of communication and political
marketing by shortening the time to produce content and targeting more
precisely,” says Othmane Zrikem, Chief Data Officer and founder of an AI
company. The shift could happen quickly in the United States, where
political advertising spending reached $9.7 billion (8.86 billion euros) in

2022, the year of the midterm elections, according to AdImpact. In fact,
the first official campaign video created by an AI was posted online on
April 25 by the Republican Party in response to Joe Biden’s candidacy
announcement. It describes a catastrophic scenario that begins with (fake) footage of the reelection of the Democratic president and continues with a fabricated sequence intended to illustrate the consequences of his reelection, from a financial crisis to China’s invasion of Taiwan.

The road is still long

In France, the major disruption that AI could bring to political practice has
not yet occurred. “Everyone is in shock over generative AI, which is a
technological update that changes the game. Data analysis software claims
to use AI, which is true, but in reality, few completed things are being
commercialized. We will probably see players emerge by the end of the
year,” says Edouard Fillias, CEO of the consulting agency JIN.
Stephane Boisson also believes that there is still a long way to go. In
Montpellier, he heads Poligma, a company of eight employees that
aggregates, analyzes, and visualizes data, mainly public data, for political
candidates or municipalities. Data overlaps that can, for example, allow an
elected official to anticipate the need for childcare places in a
neighborhood or parking spaces reserved for people with disabilities. “AI
can do very interesting things, its algorithms can imagine scenarios, it can
become a decision aid without replacing elected officials. But for now, we
are still far from it,” he says. He is convinced that “the more relevant data
is aggregated, the better the results will be when running learning models.”

He is not the only one with a company that exploits public data, the
quantities of which are increasing with the development of open data.
After working on around 1,500 electoral campaigns, including those of
Francois Hollande in 2012 and Emmanuel Macron in 2017, Arthur Muller,
Guillaume Liegey, and Vincent Pons stopped targeting elections with their
company LMP to focus on Explain, a company that markets an “AI
assistant” for companies responding to public orders.

“Every propagandist’s dream”

Among the thirty or so clients they claim are Engie, Veolia, and Bouygues
Telecom. “This allows them to understand needs by analyzing thousands
of pages of administrative documents, municipal deliberations, and
reports, which contain a lot of information,” explains Guillaume Liegey.
“It’s at the intersection of two trends: the explosion of public data that a
human cannot handle by hand, and the maturity of LMMs, the large
language models, which allow for automatic reading and writing.” They

closed a 6 million euro funding round in April (to which Minister Jean-
Noël Barrot, a shareholder in LMP since 2020, did not participate this time, according to his office and the company).

In the field of politics, it may not be for its commercial prospects that
generative AI will generate the most interest. “It will allow for much more
sophisticated and difficult-to-detect manipulation systems, on a much
larger scale than today, predicts Othmane Zrikem. More than private
companies, it will undoubtedly be state authorities, such as China, who
will be the first to use them for these purposes. It’s every propagandist’s
dream.”

Leave a comment