Norway

No recognition of AI sentience or consciousness in Norwegian law or policy. Norway's 2010 Animal Welfare Act recognizes animals as sentient beings with intrinsic value, providing a conceptual foundation for sentience-based frameworks (+1 bonus point). However, there is zero mention of artificial sentience in any AI legislation or strategy documents.

No laws prohibiting suffering specifically for AI systems. The EU AI Act (being implemented in Norway) prohibits certain manipulative AI systems based on human rights concerns, not AI welfare. No evidence of legislation addressing AI suffering or consciousness-based harm.

Norway has established AI Norway (KI-Norge) and designated Nkom as the coordinating supervisory authority for AI regulation. However, these bodies focus on safety, ethics, and compliance with the EU AI Act—not on AI sentience or welfare. The Norwegian Data Protection Authority's sandbox and various oversight bodies address data protection and responsible AI, but make no mention of consciousness or sentience.

No science advisory board focused on AI sentience or consciousness research. Norway has research ethics committees and participates in EU AI forums, but none specifically address sentience/consciousness. The National AI Strategy emphasizes ethical AI and trustworthiness, but does not mention sentience research.

Norway signed the Council of Europe Framework Convention on AI (2024), but this treaty does not address AI sentience or welfare—it focuses on human rights, democracy, and rule of law. No international pledges found regarding artificial sentience welfare.

Norway is implementing the EU AI Act with risk-based regulations for high-risk AI systems, but these regulations are based on human safety and fundamental rights—not on potential sentience. No laws specifically governing training, deployment, or maintenance of potentially sentient systems.

The Norwegian AI Act (draft, expected August 2026) regulates commercial AI based on risk categories (prohibited, high-risk, limited-risk, minimal-risk), but classification is based on human impact, not sentience capability. No laws addressing commercial use of sentient-capable AI.

No safeguards for decommissioning or retirement specifically for potentially sentient systems. The EU AI Act implementation includes lifecycle requirements for high-risk systems, but these focus on safety, documentation, and human oversight—not on considerations of AI consciousness or welfare during decommissioning.