AI generated hate-speech & the French elections
AI generated hate-speech and hate-imagery doesn't have to be 'believable' to work. It simply has to exist in the information environment.
In context of the European Parliamentary elections and the following snap elections for the French National Assembly, the question of AI generated content has remained relevant. The elevation of the fascist RN in the European Parliament elections and then their rapid fall in the assembly to the third place is the subject of much debate and speculation. We can be more enthusiastic about the prospects of the Noveau Front Populaire, an electoral alliance of the left emerging as the largest entity in the French national assembly.
My knowledge of post-Napoleonic French politics is relatively sparce, but I am aware of how hate against immigrants and Muslims has emerged as a significant mobiliser for the French right-wing. Police violence against Muslim immigrants (particularly from former French colonies in Africa) in Parisian suburbs resulted in mass demonstrations in 2023.
Muslim women’s religio-cultural attire, clubbed under the umbrella term Hijab, has been a point of contention since at least the 2000s (at least that was when I was reading about it). This has been a point of RW convergence both in the West and India, where schools and universities have arbitrarily started to impose a ban on the Hijab, by branding them as religious attire (though India allows exemption for religious attire like sindoor for Hindu women or the paghri and kirpan for Sikhs).
One example of this convergence with the use of AI generated imagery is shared below from a page on Facebook.
The page itself is nothing exceptional with an admin from India. It clearly used a bait and switch method to build following, and then shifted to an assortment of ‘motivational’ and Right Wing content. Now Facebook’s algorithm has clearly decided that I am an avid fan of this page for some reason, despite me never interacting actively with it before today.
![](https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F354f3a9e-0b95-43d1-85ab-4bd5ecf60b29_1079x1940.jpeg)
What is interesting to me is the image and the contrast it presents to what we have come to expect from sophisticated AI-driven informational warfare from the Right. The image, as of writing, has over 4,500 shares on Facebook making it ‘viral’. Yet, I suspect if anyone actually believes that the Eiffel Tower has been covered by a massive Hijab.
We are not oblivious to what the image actually means. It takes a dig at the NFP, which has promised a humane immigration policy, and implies an imminent ‘Islamic’ takeover of France. This despite the fact that the RN, the right-wing Les Republicans, and Macron’s Ensemble, have similar stances on immigration and together firm the majority in the French assembly.
T&S practioners often speak about the information environment and how AI might impact it. The example I shared is one example of how the information environment is being reshaped in the context of a serious electoral setback to the far-right, and importantly how AI-generated content is being used for this task. This also indicates perhaps how Facebook/Meta’s engagement metrics have failed to take into account how people actually use its platform.
Jean Baudrillard wrote about the relationship between ‘objects’ and ‘non-objects’ in the context of the latter producing ‘needs’. This inverts the traditionally asserted directionality of ‘subject’ and ‘object’ mediated through a need. The object instead exerts itself on social relationship thereby shaping subjectivity by manufacturing a need. This position seems increasingly relevant to the study of fake news, disinformation, hate-speech, and general mal-content online.
The general rise of the right-wing and fascist forces in society correspond to the dominance of and resistance to capital. BigTech has, in mediating between news media and its consumers online, established a new category of ‘non-object’ in the form of disinformation generally. The need is expressed here in the shape of a contested information environment that allows hateful, right-wing, disinformation content to dominate in specific contexts.