It’s no secret that Montreal has become a dynamic hub of international caliber in the field of artificial intelligence (AI). The All In international summit, which took place on September 27 and 28, also celebrates the specificity of this city which is said to be “the hub of AI in North America.
The presence of the Quebec Artificial Intelligence Institute (Mila), founded by the godfather of deep learning Yoshua Bengio, the Scale AI supercluster, funded to the tune of 337 million by the Canadian and Quebec governments and co-chaired by Hélène Desmarais, as well as the very close relationships between universities, the State and the big names in industry constitute a “tightly knitted ecosystem”. They thus make Montreal one of the centers of innovation, research and development of algorithmic machines at the forefront of what Klaus Schwab, president of the World Economic Forum, calls the “fourth industrial revolution”.
More precisely, Montreal occupies a unique niche, that of “world leader in responsible AI”. The Montreal Declaration for Responsible Development of AI (launched in 2017), the strong network of researchers on the social effects of AI, and Bengio’s leadership in calls for industry regulation and better governance of these tools are different elements that militate in favor of this honorary title.
Limits of responsible AI
And yet, the ethical approach to responsible development of AI shows several limits. Firstly, the ethical declarations which number in the hundreds around the world, and of which GAFAM are often signatories, have not changed much so far in the global trajectory of AI which amplifies social inequalities, erodes the democratic life and contributes to the worsening of the climate crisis.
Second, the voluntary approach to responsible AI, such as the voluntary code of conduct announced by the Minister of Innovation, Science and Industry, François-Philippe Champagne, as well as future Canadian laws (C -27) are tailor-made for the industry in order to reassure investors and recreate the general public’s confidence in algorithmic technologies. This approach is in line with corporate social responsibility, which serves more to give them a good image and create new markets than to effectively resolve the problem of inequalities and the ecological crisis.
Third, ethics is often limited to the reduction of algorithmic bias and questions of transparency, which ignore the central role of capitalist logic. Ethics therefore plays the role of “adviser to the prince”, and thus aims to reduce the harms of AI and give it legitimacy (social acceptability) rather than questioning its hold on States, territorial development , our social relations and our subjectivities.
A new stage of capitalism
Since the global financial crisis of 2007-2008, neoliberal, financialized and globalized capitalism has been hit by a significant crisis of legitimacy, which has been accompanied by the parallel development of cloud infrastructure, big data, machine learning , social media and smartphones as well as the consolidation of large digital platforms. This has contributed to the transition to a new stage of capitalism, in which the extraction and valorization of personal data represents a central axis of capitalist accumulation.
Data is becoming the “new oil” and AI the “new electricity” that powers our smart devices and automate a host of processes across business, supply chains, and function. public, financial transactions, etc. In other words, “algorithmic capital”, that is to say the dynamic of capitalist accumulation supported by AI, is becoming a central actor in economic life and social relations in general. Keeping us glued to our screens with addictive design techniques, constantly soliciting us, dictating our behavior through various recommendations, it sometimes even replaces human judgment through automated decision-making systems.
The AI industry here and elsewhere participates fully in the production of algorithmic capital, which today shapes a large part of our lives, often without our knowledge. Too often, the responsible AI approach ignores, or even implicitly endorses, this socio-economic system. Certainly, we highlight certain excesses of AI, piecemeal, without however recognizing the major role that capitalism and technological giants play in the creation, selection and deployment of these machines.
False solutions
Responsible AI often plays a role in solving serious problems of our time, such as the climate crisis. However, its claim to want to accelerate the ecological transition through algorithms, “green” data centers and connected objects overlooks the enormous energy cost of this global industrial infrastructure, which includes the extraction of rare metals in Congo, giga- factories in China, Amazon’s inhumane warehouses, the massive consumption of water to cool infrastructure that drives ChatGPT and the mountain of e-waste that accumulates in the Global South.
If Montreal has become the world leader in responsible AI, this ecosystem therefore constitutes the moral conscience of algorithmic capitalism. Unfortunately, this too often contributes to clearing one’s conscience and kindly supporting private actors, rather than questioning the mechanisms at the heart of this unequal, undemocratic and unsustainable economic system.
The name of the All In AI summit is also revealing: like in poker, we bet everything on AI. Everything else, the harmful effects, the automated injustices, the precariousness of work, the acceleration of the ecological crisis, the degradation of public debate caused by echo chambers and algorithmic filters, is not very important. Montreal does not stop this progress, it participates in it enthusiastically by giving it a moral veneer.