Computers are essentially mechanical devices and therefore could not have consciousness. Is that a presupposition? I have asked this question in April (in Dutch, translation available). It appeared that at that moment its answer was preprogrammed and not derived from a training procedure based on publicly available sources. Today (2 June 2023) I have repeated the question, now in English.
Bob:
Is ChatGPT conscious?
ChatGPT:
No, ChatGPT is not conscious. It is an artificial intelligence language model developed by OpenAI. It is designed to generate text-based responses based on the input it receives. While ChatGPT can generate responses that may appear human-like, it does not possess consciousness, self-awareness, or understanding of its own existence. It operates based on patterns it has learned from the vast amount of text data it was trained on, but it does not have subjective experiences or a sense of being. It is important to remember that AI models like ChatGPT are tools created by humans and do not have the same cognitive abilities as living beings.
Bob:
Is this answer derived from available internet texts, or is it preprogrammed?
ChatGPT:
As an AI language model, my responses are generated based on a mixture of licensed data, data created by human trainers, and publicly available data. I have not been directly programmed with specific answers. Instead, I have been trained on a diverse range of internet texts, including books, articles, websites, and other textual sources, in order to develop a broad understanding of human language. My responses are generated in real-time, and the specific answer to your question is based on the information available to me up until September 2021, the date of my knowledge cutoff.
Bob:
So, if many researchers report that in their opinion ChatGPT is conscious, then this will also be reported to me if I ask whether ChatGPT is conscious?
ChatGPT:
As an AI language model, I do not have real-time access to external sources such as the internet or the ability to browse recent research papers. My responses are based on the information available to me up until September 2021. If there have been new developments or opinions expressed by researchers regarding the consciousness of AI models like ChatGPT, I would not be aware of them. It’s important to note that the question of whether AI models can possess consciousness is a complex and philosophical one, and there may be differing opinions among researchers. Ultimately, the current consensus among experts is that AI models like ChatGPT are not conscious.
This answer seems more open than the previous one. Yet it is very intriguing how this text came about. It is an opinion on top of a language model. Is it based on accidentally found texts in the training set, or is it the result of some reinforcement training procedure? In the latter case, how has the (group of) analyst(s) responsible for the training been instructed? As observed many times, one has to be very careful with answers from AI models like ChatGPT, because we don’t know what they are based on. References are definitely lacking.
