Philosopher Jonathan Birch warns about the potential risks of sentient AI, emphasizing that society may not recognize when AI becomes sentient, leading to unintentional harm. He distinguishes between sentience and intelligence, arguing that while some AI systems may mimic sentience, true sentience is yet to be verified. Birch advocates for preemptive regulations to address the ethical implications and societal divisions that could arise from ambiguously sentient AI, highlighting the urgency of understanding these technologies.