Taking AI for granted means preventing digital thinking – The Conversation France

Last spring, figures as diverse as Elon Musk, Yuval Noah Harari and Steve Wozniak joined more than 1,000 “experts” to warn about the “major risks to society and humanity” posed by artificial intelligence, and a six-month break from model training to demand more powerful than GPT-4. From Le Monde to Figaro, through FranceInfo and Libération, the media have readily relayed the terms of this letter, which calls for a pause to better reaffirm the inevitability and omnipotence of future AI systems.

What is striking in the immediate media reception of this letter is the short-sightedness in the face of a process that has been theorized for almost 30 years: “digital unthinking” (or, before that, information technology). This concept of the “unthought” refers to the discursive strategies used to present technology as self-evident, often under the influence of actors whose economic or political interests it serves.

Read more: AI benefits from biased media reporting

The open letter from the “Future of Life” institute is a prime example: According to this, artificial intelligence is a powerful tool, it already exists and it is expected that it will be even more present and powerful in the future. future for the greater good of humanity.

How do you recognize the digital unthought?

Seven discursive markers should alarm you. To illustrate this “toolbox,” the open letter from Elon Musk and others, which nevertheless claims to call for a pause, presents all the discursive markers of non-thought, although it can also be applied to the very serious Villani report, in In 2018 he advocated a national and European strategy for AI:

In this type of discourse, the technical object presents itself as neutral: it is up to humanity to use it meaningfully, its mere existence serves as justification.

However, when we are told that AI poses “major risks to humanity,” isn’t that evidence that we are dealing with a technology that is not exactly neutral?

This is undoubtedly the most insidious mechanism of unthinking: the demonization of the technological object helps both to affirm its power and potential when used wisely, and to fuel the pseudo-debate, without which media interest would decline. IT, digital technology, AI are already here, there is no need to create serious historical works about them, it is enough to tell stories about entrepreneurial successes.

The unthought forms a vicious circle with the loss of political privileges…

The unthought is inextricably linked to two other processes with which it forms a truly vicious circle: the shifting of political privileges and managerialization.

Thanks to digital thinking, instruments of private actors without electoral or sovereign legitimacy even determine public access to information. An example is the X platform (formerly Twitter), which is under scrutiny by journalists because it is run by public and political figures and public institutions. When a private platform conveys a political message, we are in what is known as a “political prerogative shift.”

When private actors systematically deploy technologies, from infrastructure (cables, server farms, etc.) to software and applications, this amounts to transferring decision-making of a political nature to them. With a search engine dominating our access to information and occupying a place that would amount to a genuine public service, we are in the midst of a shift in political privilege.

We observe the same phenomenon when the French government prefers to rely on consulting firms rather than university expertise. Companies whose recommendations willingly advocate the systematic use of digital technologies and pave the way for management.

The unthought promotes the transfer of political privileges into the hands of digital actors whose tools impose themselves on citizens (excerpt from the science fiction comic MediaEntity by Simon & Émilie). MediaEntity, Simon & Emilie

…and with managerialization

Nowadays, digital tools not only allow us to manage various activities (banking, medical appointments, etc.), they have also, and above all, become indispensable for carrying out these tasks. We have no choice but to fit into the categories that these tools impose on us. It’s not always easy, for example, to make an appointment with a doctor’s assistant or to submit your tax return on paper. We call this “managerization”.

This managerialization also reflects a change. For example, the higher education access tool Parcoursup is now essential for high school students and their families. However, this instrument has a political dimension with questionable consequences, such as the exclusion of certain categories of high school graduates or the intensification of competition between training courses. In management, the tool is secondary to the activity; When it comes to management, the tool comes first: Parcoursup takes priority over the needs it is supposed to respond to.

In our daily lives, everyone obediently enters the information expected from digital consumption platforms to visit a new region, choose the menu for our dinner or meet a soul mate. When you practice a sporting activity by equipping yourself with a bracelet that processes, stores and shares a series of biometric data, this becomes the model you follow, rather than the feeling of your body, in a kind of “self-management”. .

Properly identified and profiled through our instruments, we contribute unreservedly to the economic profits of companies whose income is largely tax-free… and therefore to democratic power, already shaken by the shift in political prerogatives.

Digital productivity and digital management (Sociopolitical issues of the digital, Dominique Boullier).

Criticize… and act

Digital technology is not the sometimes hellish and sometimes bright future that its unthinkers promise us: it is simply a category to designate a series of technical objects and socio-technical devices that need to be questioned and discussed in terms of their political and social impact.

As the unthought draws our attention to AI, perhaps we need more new tools (which may have AI in them) to better organize the (digital) expression of our intelligence in the face of problems that require unprecedented collective decisions. Climate, democracy, environment, health, education, coexistence: there is no shortage of challenges.

With this in mind, we invite you to discover the new version of the contribution-based web navigation service Needle. This radically different proposal for accessing and sharing digital content is based on the concept of unthought and is based on collective intelligence. Needle is a networking platform that realizes the hope of a digital environment rich in connectivity and curious exploration for all, rather than a network of straight lines through which artificial intelligences are supposed to tell us which documents to consult.

This technology is now supported by a start-up, proof that it is possible to design concrete proposals that take into account the necessary criticism of the place of technology in our societies.