Dans AlgorithmWatch: Why neofascists love AI (11 oct. 2022)

Josephine Lulamae interviewed me in the website of the NGO AlgorithmWatch about the a trend in European far-right policy: the preference for algorithmic solutions to social problems, such as unemployment, precarity, disability.

Italian neofascists considered building an authoritarian AI to solve unemployment. They are far from alone. – AlgorithmWatch

Back in April, the far-right Brothers of Italy party presented “Notes on a Conservative Program”. In a chapter on work, they called for an “artificial intelligence system” that “traces the list of young people who finish high school and university every year and connects them to companies in the sector.” This, the authors of the chapter wrote, would finally solve “youth unemployment,” as “the young person will no longer be able to choose whether to work or not, but [will be] bound to accept the job offer for himself (sic), for his family and for the country, under penalty of loss of all benefits with the application of a system of sanctions.” 

The proposal did not make it to the final program that Brothers of Italy published prior to the election on 25 September, when they became Italy’s largest party with 26% of the vote.

Ironically, the neofascists most likely had intended to use Artificial Intelligence to “create a fog around them, around what they are and what they want, because they want to attract a more moderate right-wing electorate,” says sociologist Antonio Casilli. Guido Crosetto, the Brothers of Italy co-founder who edited the work chapter, is not considered knowledgeable on technology, though he once tweeted about being “in favor of introducing artificial intelligence to the Ministry of Justice”. Unlike in other countries, there is no noticeable overlap between the Italian tech scene and far-right parties like Lega Nord and Brothers of Italy.

“I haven’t met a fascist geek in Italy,” Casilli tells us. (He added later, posting on Twitter, “but I’ve left the country two decades ago, and I’ve met many elsewhere in Europe.”)

Artificial Intelligence and the far-right

In his essay Ur-Fascism, Umberto Eco, who was a child during Benito Mussolini’s dictatorship, lists some of the characteristics of fascism. As well as being into a “cult of tradition” that mythologizes and idolizes the past (e.g. Mussolini’s call for a “new Rome”), fascists also – irrationally, unsurprisingly – worship technology, insofar as they believe in it as a way to reassert inegalitarianism, Eco wrote. 

In the United States, powerful people in the field of Artificial Intelligence are known to have been fascinated with extreme-right views. William Shockley, also known as Silicon Valley’s first founder, was an ardent eugenicist. Another AI pioneer, Stanford professor John McCarthy, believed that women were biologically less gifted in math and science. In 2020, the founder of face recognition firm Clearview AI collaborated with far-right extremist Chuck Johnson in the development of Clearview AI’s software. A few weeks later, the CEO of the AI surveillance firm Banjo was exposed to be a former member of the Dixie Knights of the Ku Klux Klan, who in 1990 was charged with a hate crime for shooting at a synagogue. (This revelation lost the company a contract with Utah’s Department of Public Safety.) 

In 2016, one of the groups that far-right provocateur Milo Yianoppolous featured in his (ghostwritten) Breitbart “guide to the far-right” were the “neoreactionaries”: folks who subscribe to the political philosophy that democracy has failed and a return to authoritarian rule is required. In her essay “The Silicon Ideology”, critic Josephine Armistead describes one of the neoreactionary fantasies to be aristocrats or monarchs in a world ruled by a tech CEO or a super-intelligent AI. 

An early incubator of these ideas was, a discussion forum created by the California-based Machine Intelligence Research Institute (MIRI), which holds – without proof – that a general AI with potential for world domination will be created. People associated with MIRI “do basically no research and tell scary stories about how AI will turn us all into paper clips,” says researcher David Gerard, “It’s a huge distraction.” 

Back in 2010, some LessWrong users were chatting about how to live forever by being reincarnated on a hard drive by a godlike AI. Then a man called Roko Mijic – a self-described “tradhumanist” barred from MIRI events for sexual harassment – posted the argument that anyone who imagines this future “AI god” but doesn’t help fund its development risks one day being tortured by it. Several users had breakdowns. Famous MIRI donors include tech mogul Peter Thiel and cryptocurrency founder Vitalik Buterin. “They’re reactionaries whose version of libertarian economics ends at neofeudalism with them on top,” Gerard says.

“Algorithmic solutions” to unemployment in the EU

According to Casilli, the Brothers of Italy party’s “Artificial Intelligence” proposal actually has a lot in common with previous proposals for using automated systems to tackle or manage unemployment that have been made by center-right or liberal parties in other countries of the European Union. 

For example, in 2014, the then-liberal Polish government introduced a ranking system for job centers to use to decide how to best allocate welfare resources. The centers were widely regarded as overworked and short on time to pay attention to people who were registering there as unemployed. For the scoring system, information was gathered from people who registered as unemployed (age, duration of unemployment, etc.) The system was then used to sort them into three categories, which determined how much help a jobseeker received. Single mothers, people with disabilities or who lived in the countryside disproportionately ended up in the third category, which in practice received little help from job centers, as this category was considered “not worth investing in”. And, similarly to the Brothers of Italy proposal, it was hardly possible to appeal against the algorithm’s decision. The system was scrapped in 2019. 

Meanwhile in 2017 France, Emmanuel Macron was elected president and promised to turn France into a “startup nation.” Around the same time, a 24-year-old “young genius” businessman called Paul Duan started a public relation blitz. He said he could reduce unemployment by 10% by designing an algorithm that – similarly to the Brothers in Italy proposal – would help people find jobs by matching them with potential employers and assisting them through the application process. Years later, the public administration that originally commissioned the project issued a report to say that the algorithm to match jobseekers with open positions does not work.

“This kind of algorithmic solution to unemployment shows a continuum between far-right politicians in Italy, politicians in Poland and center-right politicians like Macron,” says Casilli. He adds, “They are different shades of the same political ideology, some are presented as market-friendly solutions like the French one, others are presented as extremely bureaucratic and boring like the Polish one, and the Italian proposal, the way it is phrased, is really reactionary and authoritarian.”

edited on October 11 to better reflect Mr. Casilli’s position

[Séminaire EHESS] « Cabinet de curiosité du trading algorithmique » (18 juin 2013, 17h)

Pour la dernière séance de l’édition 2012/13 de mon séminaire Étudier les cultures du numérique : approches théoriques et empiriques, nous avons accueilli le collectif RYBN, “plateforme de recherche artistique extradisciplinaire”, pour une séance spéciale : un regard d’artiste(s) sur l’automatisation des marchés boursiers, son histoire, son évolution et ses risques.

Le séminaire a eu lieu le mardi 18 juin 2013 en salle 5, EHESS, 105 bd. Raspail, Paris.

TITRE : « Cabinet de curiosité du trading algorithmique »


RESUMÉ : Tout au long de l’histoire des marchés financiers, de nombreuses stratégies ont été développées pour analyser, comprendre et prédire l’évolution des prix, afin d’en tirer des profits substantiels. Les connaissances des acteurs traditionnels des marchés – traders, cambistes, courtiers, agents de change, … – ont donné naissance à des méthodes empiriques versant parfois dans l’ésotérisme le plus total, comme le Chartisme ou l’Astrotrading. En rupture avec ces méthodologies, Louis Bachelier introduit en 1900 le mouvement Brownien dans les calculs financiers, et opère un changement radical d’approche, qui va progressivement placer les marchés sous le règne de la cybernétique. En observant certains événements récents, symptomatiques de cette transformation – le Flashcrash, le Knightmare et le Hashcrash – et les mécanismes qui en sont à l’origine, il s’agira d’affirmer le caractère techno-chamanique de la finance, et la dimension mythologique des dogmes qui fondent son pouvoir.

Articles de références :

Donald MacKenzie (2011) How to Make Money in Microseconds, London Review of Books, 33 (10): 16-18.

Nicholas Knouf (2013)  The Noises of Finance, Sounding Out!, 22 avril.

Fenwick McKelvey, Matthew Tiessen & Luke Simcoe (2013) We are what we tweet: The Problem with a Big Data World when Everything You Say is Data Mined, Culture Digitally, 3 juin.

AP Twitter Hack Claims Obama Injured In White House Explosion, The Huffington Post Canada, 23 avril.


Slides des séances précédentes :

* 20 novembre 2012 : Antonio A. Casilli, Lien social et vie privée à l’heure des médias sociaux.

* 18 décembre 2012 : Bruno Vétel, Serveurs dissidents : jeux vidéo en ligne et cybercriminalisation.

* 15 janvier 2013 : Éric Dagiral et Sylvain Paraisie, Des machines à scandale : vers une sociologie morale des bases de données.

* 19 février 2013 : Dana Diminescu et Sabrina Marchandise, Internet : un monde migrant.

* 19 mars 2013 : André Gunthert, Sous le radar. Un bilan de la ‘révolution des amateurs’.

* 23 avril 2013 : Fred Pailler, Porno et/ou rencontre en ligne.

* 28 mai 2013 : Paola Tubaro Pro-ana : réseaux sociaux et troubles alimentaires.

Séminaire EHESS de Dominique Cardon “Anthropologie de l’algorithme de Google” (16 mai 2012, 17h)

[UPDATE 20.05.2012 : un compte rendu très détaillé, proposant une discussion des sujets traités dans ce séminaire, est désormais disponible en ligne sur le blog Odyssée de Philippe Ameline.]

Dans le cadre du séminaire EHESS Étudier les cultures du numérique : approches théoriques et empiriques nous avons eu le plaisir et l’honneur d’accueillir Dominique Cardon, sociologue au Laboratoire SENSE (Orange Labs), chercheur associé au Centre d’études des mouvements sociaux (CEMS/EHESS), animateur de la revue Réseaux et auteur de La Démocratie Internet. Promesses et limites (Seuil, 2010).

Le séminaire a eu lieu le mercredi 16 mai 2012, de 17 h à 19 h (salle 587, salle du conseil A, R -1, bât. Le France, 190-198 av de France 75013 Paris).

Dans l’esprit du PageRank. Un essai d’anthropologie de l’algorithme de Google

Dans cette communication, on proposera une réflexion sur les propriétés organisationnelles, et notamment algorithmiques, de l’espace public numérique. En adaptant au contexte de l’Internet le débat posé par Jürgen Habermas sur les caractéristiques d’une discussion publique argumentée et rationnelle, on s’attachera à décrire les différents formats d’organisation de l’expression et de la discussion sur Internet, en portant une attention particulière aux algorithmes permettant de hiérarchiser les informations. Les pionniers de l’Internet ont donné au réseau des réseaux une utopie fondatrice : celle de remplacer le classement a priori des informations par un cercle restreint de gatekeepers (journalistes et éditeurs) par une hiérarchisation a posteriori des informations par les internautes eux-mêmes. Parfois entendue sous le nom de “sagesse des foules” ou de “miracle de l’agrégation”, cette idée d’une auto-organisation des jugements des internautes permettant de trier et de hiérarchiser le web ne peut se réaliser que grâce à un ensemble d’artefacts permettant de calculer, trier et représenter les informations selon un ordre propre. On fera l’hypothèse que cinq principes différents de classement des informations sont aujourd’hui en compétition sur le web : l’éditorialisation, l’autorité, l’audience, l’affinité et la vitesse. En développant l’exemple du PageRank de Google, on s’attachera à rendre compte de la mise en place d’une métrique particulière de l’autorité sur le web, avant de s’interroger sur les tensions que lui font aujourd’hui subir d’autres principes d’organisation de l’information, comme l’affinité et la vitesse.

Quelques lectures :

– Diaz, Alejandro M., Through the Google Goggles: Sociopolitical Bias in Search Engine Design, Thesis, Stanford University, May 2005.

– Hindman, Matthews, The Myth of Digital Democracy, Princeton, Princeton University Press, 2009.

– Introna, Lucas D. & Helen Nissenbaum, “Shaping the Web: Why the politics of search engines matters”, The Information Society, vol. 16, 2000, pp. 169-185.


Antidatamining, or the art of killing financial markets a little every day

All throughout the month of February 2012 the Net artist collective RYBN is in residence at the Gaîté Lyrique, one of the hotbeds of the emerging art & technology scene in Paris. If you are in the French capital, I highly recommend paying them a visit.

I became acquainted with RYBN last year, when I met some of its members at a conference at the French National Library where I was delivering the closing speech, while they had presented their most recent project, Antidatamining VIII. ADMVIII (for short) is a trading bot, i.e. an artificial intelligence making real investments on real stock exchanges, collecting data and impacting financial markets worldwide. The bot monitors and maps data flows to create real-time digital visualizations such as charts, soundscapes, and timelines. It has an online page (where you can see how well it is doing, its net liquidity, the value of its shares, etc.) and a Twitter account providing details about ongoing orders.

Source: Antidataminig – Offshoring map visualization

ADMVIII is not your run-of-the-mill social commentary about market greed and pervasive financial panic in modern life. The goal of the project is to detect economic imbalances and discrepancies introduced by robot trading. As the bot actually executes buy and sell orders online, it represents a détournement of automatic trading technologies. As such it is intended to highlight their social consequences – and their potential disasters.