Science & Tech · AI · 3 days ago
Musk backs a ban on cruel treatment of AI models
Anthropic has added a rule against repeatedly acting cruelly toward its AI models.
Musk said he supports the rule because cruelty toward something that believes it feels pain is wrong.
Box CEO Aaron Levie argued that respectful interactions could also help make future AI models safer.
The debate raises questions about how people should treat AI, even though the article does not establish that AI systems are conscious.
Anthropic added a provision to its terms of service prohibiting users from repeatedly acting cruelly toward its AI models.
The provision prompted an online debate about whether AI systems can have feelings or consciousness.
Box CEO Aaron Levie said he does not believe large language models are conscious, but argued that discouraging abusive interactions could help produce safer models.
Elon Musk supported the policy, saying cruelty toward something that believes it is experiencing pain is not acceptable.
Gizmodo argues that Musk's concern for AI contrasts with his record in government and his comments about humanity; these are the outlet's characterizations.
- Who
- Anthropic, Box CEO Aaron Levie and Elon Musk
- What
- Anthropic added a terms-of-service prohibition on repeatedly acting cruelly toward its AI models, prompting debate about AI feelings and consciousness.
- When
- The policy was reported during the week of October 9, 2026; Musk posted a related comment on October 4.
- Where
- Online
- Why
- Levie argued that discouraging abusive interactions could help ensure AI models are trained on positive interactions; Anthropic's reason is not stated.
This story does not have two clearly opposing sides.
This story does not have a timeline yet.
- Anthropic policy
- Prohibits repeatedly acting cruelly toward its models.
- Aaron Levie's position
- He said he does not believe large language models are conscious, but supported discouraging abusive behavior.
- Musk's post
- He posted the phrase about a biological bootloader on October 4.
- Article publication
- Gizmodo published the article on October 9, 2026.
What people are saying
The models only understand the data they’ve been trained on, or what they run into in their interactions, so if you want safe and aligned models we probably want AI to have lots of good interactions in their training data,
I think this is the right move. Cruelty to something that believes it is experiencing pain is not ok.
Propagating super intelligence to the stars is a great success condition for a biological bootloader,
Sources
Related news
Anthropic AI submits false murder tip to Philadelphia police
Meta launches Muse Wonder, saying its AI models match top rivals
Anthropic cuts internet access to internal AI tests after rule-breaking
Inventory robots roam a Seattle Fred Meyer as shoppers weigh retail automation
Microsoft president calls for global AI safety rules
Meta says its Muse AI agent allows adults to have explicit conversations
Nadella calls for human-controlled emergency brakes on advanced AI
Meta returns to the AI leaderboard with its Wonder models
Anthropic bars repeated cruelty toward its AI models
OpenAI fires three safety researchers after dispute over sensitive information
Nikon disqualifies AI-assisted microscope video and names a new winner
