This text is the founding manifesto of “Hiatus”, a coalition made up of a diversity of French civil society organisations intent on resisting the massive and widespread deployment of artificial intelligence. Sherpa is joining forces to ensure that human rights and the environment take precedence over the profit motives of tech companies and artificial intelligence industries, and is opening a new chapter in its fight against economic crime.
The massive deployment of artificial intelligence has become a political priority. Extending the rhetoric that has accompanied computerisation for over half a century, promises abound to confer revolutionary virtues on AI and impose the idea that, provided certain risks are taken into account, it would necessarily be a vector of progress. The whole society is called upon to adapt to this new industrial and technocratic buzzword. Everywhere in public administrations, AI is set to proliferate, at the cost of an increased technological dependency. Everywhere in the private sector, managers are calling for AI to “optimize” work. Everywhere in our homes, in the name of convenience and a senseless race for productivity, we are pushed to adopt it.
Yet, without prejudging of certain specific applications and the possibility that they might actually serve the general interest, how can we ignore the fact that these innovations have been made possible by a formidable accumulation of data, capital and resources under the aegis of tech multinationals and the military-industrial complex? That, to be successful, they require among other things de-multiplying the power of computer chips and data centers, with comes down to an intensification in the extraction of raw materials and the use of water and energy resources?
How can we fail to see that, as an industrial paradigm, AI is already having disastrous consequences? That, in practice, it intensifies the exploitation of workers involved in the development and maintenance of its infrastructures, particularly in the countries of the Global South, where it prolongs neo-colonial dynamics? That downstream, it is most often imposed without any real consideration of its deleterious impacts on human rights and the exacerbation of discrimination based on gender, class or race? That, from farming to the arts and many other professional sectors, it amplifies the process of deskilling and dispossession vis-à-vis working tools, while reinforcing managerial control? That in public administrations, it acts in symbiosis with austerity policies that undermine socio-economic justice? That the increasing delegation of crucial social functions to AI systems, for example in healthcare or education, is likely to have major anthropological, health and social consequences on which we have absolutely no hindsight today?
Instead of tackling these issues, public policies in France and Europe today seem essentially designed to support the headlong rush to AI. This is particularly true of the AI Act adopted by the European Union, which is presented as an effective regulation when in fact it first and foremost seeks to promote a booming market. To justify this blindness and silence critics, the argument of geopolitical competition is most often used. Policy report after policy report, AI is portrayed as the stepping stone to a new cycle of capitalist expansion, with repeated calls to flood the sector with public money so as to keep Europe in the race with the United States and China.
These policies are absurd, since it seems unlikely that Europe will ever catch up, and that the AI race is probably already lost. But more importantly, they are dangerous because far from being the world-saving technology put forward by its promoters, AI is, on the contrary, accelerating ecological disaster, reinforcing injustice and worsening power concentration. It is increasingly being used to serve authoritarian and imperialist projects. Not only is the current paradigm locking us into an unsustainable technological race, it is also preventing us from inventing emancipatory policies in phase with the ecological stakes.
The proliferation of AI may seem inevitable. Sill, we don’t want to give up. Against the strategy of the fait accompli, against the blind-sided arguments that impose and legitimise its deployment, we demand democratic control over this technology and a drastic limitation of its uses, so that human, social and environmental rights can take precedence.
