Anthropic protects the welfare of possible future AIs. Present-day humans, not so much

What if large language models are actually the Singularity? What if chatbots are intelligent beings we have moral obligations to — and not just statistical slop to predict the next token in sequenc…