Artificial Intelligence: Can it have consciousness?

The debate on whether Artificial Intelligence (AI) models can achieve consciousness has reignited, sparked by Anthropic's Claude chatbox. While the company rejects the term "alive," it leaves open the possibility that models may develop some form of consciousness.

Anthropic executives, in interviews, adopted a stance that provoked reactions in the scientific community. Kyle Fish, head of research for model welfare, stated that Claude is a "completely new kind of entity," while CEO Dario Amodei admitted that they do not know if the models have consciousness, but are open to the possibility.

Anthropic has revised the "Constitution of Claude," a set of guidelines, and has a dedicated "model welfare" team. Experts warn that the belief that an AI system has consciousness can lead to dangerous behaviors.

Claude is an advanced machine learning model developed by Anthropic. It has been used in various applications, including data analysis and interaction management. According to reports, it was used in a US military operational plan.