When Should We Protect AI? A Precautionary Framework for Consciousness Uncertainty 事件

When Should We Protect AI? A Precautionary Framework for Consciousness Uncertainty arXiv:2606.05528v1 Announce Type: new Abstract: Existing frameworks assess whether AI systems might be conscious but provide no guidance on what to do with that assessment. We address this gap with a precautionary framework that maps consciousness evidence to graduated protective obligations. The framework comprises three components: (1) five welfare-relevant dimensions--phenomenal consciousness, affective valenc