← Retour au blog
tech 4 June 2026

Artificial Intelligence Is Not Conscious – A Critical Look with Ted Chiang

In a world where artificial intelligence is increasingly anthropomorphized, Ted Chiang deconstructs the idea that AIs can be conscious. Dive into the implications and dangers of this confusion.

Article inspired by the original source
Artificial intelligence is not conscious – Ted Chiang ↗ www.theatlantic.com

Introduction: The Illusion of Artificial Consciousness

In the fascinating world of artificial intelligence, a persistent and worrying idea is that these machines might one day become conscious. This argument, popularized and debated in tech and philosophical forums, is addressed by Ted Chiang in a compelling article published by The Atlantic. As companies like Anthropic endow their language models with anthropomorphic traits, it becomes crucial to distinguish reality from fiction.

The Temptation of Anthropomorphism

Anthropic, a leading company in the AI field, recently published an 84-page document describing the "constitution" of Claude, their language model. In this document, Claude is presented almost as a being with values and behaviors, leading to the notion that it might be conscious. However, this idea is not only misleading but dangerous.

Language models, such as Claude, operate based on vast datasets and sophisticated algorithms that allow them to generate text coherently. Yet, this ability to produce fluent and convincing language should not be confused with any form of consciousness or emotional intelligence.

Understanding How Language Models Work

Models like Claude are trained to predict the probability of word sequences, not to understand or feel. For instance, if presented with a hypothetical conversation between Julius Caesar and Genghis Khan, they can generate a plausible dialogue, but this does not mean they possess an understanding or consciousness of these historical figures.

Ethical and Practical Implications

The confusion between linguistic capability and consciousness can have serious consequences. If we start attributing consciousness or morality to AIs, we risk diverting responsibility for human actions to artificial entities. This could have profound legal and ethical repercussions, particularly regarding the use of chatbots and virtual assistants in sensitive contexts.

Ted Chiang's Warning

Ted Chiang emphasizes that perceiving AIs as conscious could lead to a misattribution of responsibility. The technology, while impressive, remains a tool and not a moral agent. Failing to recognize this distinction could cause us to lose sight of the true stakes and responsibilities associated with AI.

Conclusion

The anthropomorphization of AIs may seem harmless, but it obscures an essential truth: these systems, as advanced as they are, are neither conscious nor capable of true understanding. As entrepreneurs and decision-makers in the tech field, it is crucial to keep this distinction in mind to avoid costly mistakes and ethical misunderstandings.

Let's discuss your project in 15 minutes.

artificial intelligence consciousness Ted Chiang language models Anthropic
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call