I believe current AI models can think and understand – Part 1 of 2 (blog.andymasley.com)

In a thought-provoking exploration, a prominent philosopher argues that current AI models possess the capacity for "thinking" and "understanding," albeit not in the same way humans do. He delineates between two types of consciousness: phenomenal consciousness, which refers to subjective experience, and access consciousness, characterized by the ability to process and share information within the mind. The philosopher posits that AI may exhibit access consciousness, allowing it to perform tasks that resemble thinking without any inner experience akin to human consciousness. This distinction prompts a reevaluation of how we define thinking and understanding in both humans and AI. The significance of this perspective lies in the broader implications for the AI/ML community. If AI can indeed "think" by the philosopher's criteria, it challenges long-held beliefs about the limitations of AI technology, particularly the idea that it cannot innovate or engage in complex reasoning beyond programmed capabilities. By reframing the debate around AI's capabilities and the potential dangers of anthropomorphizing these systems, the philosopher encourages a more nuanced understanding of AI's functions. This discussion may reshuffle assumptions about AI's role in society, from mere tools to entities that could contribute meaningfully to innovations and complex problem-solving in a manner previously thought exclusive to humans.