Artificial intelligence technologies are rapidly transforming every aspect of life and creating a new ecosystem. In particular, generative artificial intelligence technologies such as large language models (LLMs) have been adopted and used by a large-scale audience at an unprecedented speed. While the opportunities offered by artificial intelligence are being explored across economies, the risks associated with these technologies have also begun to emerge.
A serious epistemic crisis has become increasingly visible, particularly in relation to LLMs. The misinterpretation of the nature of a machine that imitates human cognitive processes, as well as its potential replacement of human judgment, has created a new area of concern. While human beings relate to the world through truth, values, and responsibility, artificial intelligence has neither a claim to a relationship with truth nor a sense of responsibility. Human judgment is the product of a multidimensional process extending from perception to causality and from memory to values, whereas the judgment of LLMs is essentially based on selecting the next word within a high-dimensional probability space.
Artificial intelligence systems operate through data and algorithms, functioning not within a space of truth but within a space of probabilities. Problems such as bias and hallucination also emerge from this structure. The fact that artificial intelligence can act as an agent despite lacking consciousness, responsibility, and intention raises significant ethical concerns. The replacement of our relationship with truth by AI-generated outputs, and the emergence of an object without responsibility as an apparent epistemic subject, constitute one of the central challenges of the current age.
Therefore, the problem cannot be resolved simply by improving the training of models or increasing their accuracy; it is also a structural issue. No matter how convincing these systems become, they are not epistemic subjects. Treating them as such may undermine individual judgment and institutional responsibility. In the age of generative artificial intelligence, the fundamental question is who retains judgment and who is the agent. Rethinking the epistemic, institutional, and ethical boundaries necessary to preserve human judgment is therefore essential for building a fair, authentic, and human-centered future.
