It is now everywhere. In science, medicine, photography, cinema, the media, art or sport… Generative artificial intelligence – and its new star ChatGPT – is gradually disrupting our societies. The National Assembly is not spared from this new digital “revolution”. But what do the deputies think? Between distrust and fascination, they try to adapt to the emergence of these new tools.
“AI is like porn”
“The subject does not inspire me. It’s still too far away for a provincial deputy who plays on proximity to get by,” quips Eric Pauget. Like the elected LR from Alpes-Maritimes, most members of the Assembly do not really use ChatGPT, the conversational robot developed by OpenAI, nor other forms of AI in their practices. “I am a little old school. An MP speaks and writes, summarizes Constance Le Grip. Personally, I don’t use ChatGPT at all. And I’m not sure that this kind of AI will be that useful to us…”
The elected representative of Renaissance des Hauts-de-Seine, who released a report on the supervision of AI at the European level, is however curious. “I try to document myself because I am aware that this is not just another step in the evolution of digital practices; it’s a new revolution. I also suspect my colleagues of getting hold of it from time to time without telling me…” she laughs.
A left-wing elected official makes the same remark: “I don’t use it, but my teams do a lot. Some interns will produce a note that is a bit dirty in terms of writing, they put it in ChatGPT to come out with something good,” he smiles. A use out of sight which does not surprise the socialist Hervé Saulignac: “There is, even today, a somewhat shameful use of AI in politics, it remains taboo. It’s like porn, no one watches it…”
“We will not be elected thanks to ChatGPT”
Thanks to him, ChatGPT had its moment of glory in the hemicycle of the Palais Bourbon. In April 2023, the elected PS asked the government a question on the rise of artificial intelligence. Before revealing the deception: “Mr. Minister, you have just responded to the 578th deputy at this moment, since the question I have just asked you was entirely written by ChatGPT”. The start of a major replacement of deputies to create the law? Not really, because apart from a few highlights, like this question or the writing of an amendment by the AI, its use mainly comes down to saving time.
“ChatGPT, I happily use it with my teams. For translation into English or to speed up fundraising before speaking. Presidents do have feathers, what’s the difference? », asks Renaissance MP Anne Genetet. “When we lack ideas, it can help, open up avenues for reflection. You have to be aware of your limits, particularly on the question of sources, but it’s a fantastic tool,” she continues.
The rebellious MP Antoine Léaument also uses several AI on a daily basis. “It saves a phenomenal amount of time on time-consuming tasks such as writing descriptive texts under our videos, publishing posts on the networks…”, lists the former “digital gentleman” of Jean-Luc Mélenchon… who works on elsewhere to the design of a 100% rebellious tool, nicknamed Robotspierre (yes, like the revolutionary).
Others cite the use of ChatGPT to synthesize data. “We give him something to eat and he quickly gives us a summary. I tried with a big thesis of 600 pages, it’s rather practical, recognizes Arthur Delaporte. But AI has not yet changed the way we work, it remains daily tinkering, nuance the PS MP. It can help for a speech frame, but not for a program. We will not be elected tomorrow thanks to ChatGPT.”
Between bursts and fears
Image or video generators are also used by a select few, more for fun than necessity. This is the case of Hervé Salignac, who performed with Charles de Gaulle, Coluche and Simone Veil to send their happy new year wishes. “I used AI but it was a homemade video, a bit dirty for fun. We did not try to do something that could become misleading,” he adds. Because obviously, the ocean of new possibilities offered by AI arouses many fears in the political world. Last year, Renaissance MP Éric Bothorel filed a complaint against ChatGPT with the Cnil* to denounce the false information provided by the robot on his bio.
“It was more of a crash test than a condemnation of the AI. At the time, there was no awareness of the limits of the tool,” he says. “AI will enable wonderful things, but the political world will not escape misuse. We will move into a world where the true will coexist with false things which will be similar to the true, like deep fake “. Public officials fear that these fake videos, capable of making a political figure say anything, and whose realism is progressing at breakneck speed, will wreak havoc in the electoral campaign. Last January, Renaissance spokesperson Loïc Signor published on his X account a video generated by AI by Marine Le Pen speaking in Russian, to denounce the supposed links of the RN with Moscow. Proof of the political concern around the excesses of AI, his initiative had been condemned even within the macronie.
* National Commission for Information Technology and Liberties.