☰ Aymeric Blot

Generative AI is an incredible technology that has been found genuinely useful in many applications.

But it is not magical; whilst it may appear fluent and convincing, at its core lie statistical patterns, not knowledge or causal understanding[1]. It is used in creative contexts, but has no inherent understanding of correctness or originality[2]. AI models are inscrutable, yet they are increasingly used to drive decisions with real and tragic consequences[3].

The impact of generative and especially agentic AI on the environment is undeniable and devastating[4]. Despite compounding training and operating costs and no path to profitability, tech companies have no market incentive to slow down[5]. Much of the human-created material used for training was collected without meaningful consent, compensation, or attribution[6]. AI as a technology is largely unregulated, and fundamental ethical questions such as bias and alignment show no sign of ever being answered[7].

Just as social media and infinite scrolling erode focus, memory, and mental well-being, growing dependence on AI-generated content undermines judgement, creativity, and cognitive well-being[8]. In education and academia, shortcuts that bypass the effort required to learn and acquire expertise threaten the very process through which future generations of independent thinkers and researchers are formed[9]. In science and software engineering, the drive towards automated conclusions and immediate results sidelines foundational work and meaningful new research questions[10].

Overall, although generative AI has clear merits, the current push towards its unbridled and accelerated integration into every aspect of work, education, and society is unsustainable and profoundly harmful to humanity, and its impact on nature is indisputable and unforgivable.