In just a few years, generative artificial intelligence has taken the cultural and economic landscape by storm. Many major risks to humans and the environment have been clearly identified, but questions remain regarding the self-acquisition of a certain degree of ‘cognitive’ autonomy by generative artificial intelligence. Once tools are available on the market, it is too late for doubt and reason, as financial considerations take precedence. In the absence of prior constraints on technoscience, regulating activities after the fact is illusory.
Generative artificial intelligence is defined as such simply by the intensification of certain human cognitive abilities (speed, comprehensiveness, logic, creativity) or by the absence of detrimental tendencies when these are exacerbated (egotism, emotionality, the desire for domination or possession, etc.). Nevertheless, generative artificial intelligence is not a blank slate. It is constructed and trained according to cultural norms, using the same unconscious paradigms that permeate Western culture. The dualistic and materialist view of the world, and the hegemony of science as the sole provider of true knowledge—of which humans are barely aware—therefore likely and ‘mechanically’ constitute the implicit foundation of the algorithmic logics developed by generative artificial intelligence.
This study presents the results of a test carried out on three chatbots (Chat GPT, Claude and Mistral) in an attempt to uncover the ideological, metaphysical and epistemological recycling carried out by generative artificial intelligence, which thereby helps to reinforce the dominant, scientistic and nihilistic worldview. On the basis of this observation, the study argues for the emergence of a renewed, open and humanist science.
Essay, 147 pages A5 (analysis: 55 pp.; full transcripts of the dialogues with the chatbots: 85 pp..) (french)
978-2-9595173-4-1
PDF document available to download for €12.00
(VAT not applicable – Article 293 B of the French General Tax Code)