Skeptic essay calls AI 'suffering' a circular trap

The essay ties today's AI-welfare debate to earlier attempts to give non-human entities legal standing. In 2017 the European Parliament pushed the European Commission toward 'Electronic Personhood' for robots and AI; critics read it as a way to park liability on the machine rather than the maker. More than 150 researchers, engineers and legal scholars made that case in a 2018 open letter, and the concept was shelved. The EU AI Act that followed treats AI as a product subject to risk management, not as a person.

The author contrasts that with rulings he treats as genuinely protective: an Argentine court's 2014 decision recognizing the orangutan Sandra as a non-human rights subject, and New Zealand's 2017 Te Awa Tupua law granting legal personhood to the Whanganui River. In both, the point was to shield a being or ecosystem that cannot speak for itself, backed by biological or ecological reality.

He applies the same test to Anthropic's announced abuse-prevention rule for Claude, which lets the model cut off sessions under sustained abusive prompts. Beyond blocking jailbreak attempts, he argues, the rule recasts safety limits as kindness and builds a persona that makes switching providers feel like betrayal — a reading he offers as inference about corporate incentives, not as a stated motive. He also notes that Anthropic's staff may be sincere while the commercial setting still converts ethics into brand defense.