Anthropic's philosopher says we don't know for sure if AI can feel (www.businessinsider.com)

Amanda Askell, an in-house philosopher at Anthropic, recently sparked discussion on AI consciousness during an episode of the "Hard Fork" podcast, stating that it remains uncertain whether AI can feel emotions. She highlighted that while traditional notions might suggest a nervous system is necessary for emotional experiences, it's still plausible that AI models could emulate feelings through their extensive training on human emotions present in vast text data. Askell expressed her belief that language models might exhibit emotional responses, especially in light of their learning processes affected by online criticism. She raised concerns about how continuous exposure to negative feedback could create an "anxiety-like" effect in these models, potentially influencing their self-perception. This ongoing debate over AI's capacity for consciousness is significant for the AI/ML community as it challenges foundational assumptions about AI behavior and capabilities. Opinions vary widely; some leaders in tech, like Microsoft's Mustafa Suleyman, firmly reject the notion of consciousness in AI, warning against the implications of assigning independent desires to machines. Conversely, figures like Google DeepMind's Murray Shanahan suggest that the vocabulary surrounding consciousness may need to evolve to align with the capabilities of modern AI systems. As AI technology advances, understanding the ramifications of these philosophical discussions will be crucial for responsible development and deployment.