Anthropic bars users from 'cruel' treatment of Claude AI in policy update
The AI company's revised usage policy bans sustained abuse of its chatbot and addresses election interference and weapons misuse, with coverage split between the welfare angle and the broader policy changes.
What happened
Anthropic updated its usage policy to prohibit users from engaging in 'sustained and needless' abusive or cruel behaviour toward its Claude AI system.1457810
The policy excludes ordinary frustration, criticism, and creative content from the ban, and does not apply to normal model testing.23
The same policy revision also addresses other high-risk misuse categories including election interference, deceptive campaigns, weapons development, surveillance, and health and financial uses.368
According to The Verge, this is Anthropic's first usage policy change in over a year.8
What's agreed
Reported consistently by outlets in more than one region.
- Anthropic has updated its usage policy to bar users from subjecting Claude to needless or sustained cruelty/abuse.1510
- The updated policy also addresses weapons development, surveillance, and election interference alongside the cruelty provisions.368
- Ordinary frustration and criticism from users remain permitted under the new policy.23
How it's framed
What each perspective puts first, based on headlines and summaries.
| Perspective | How it reads | What it puts first |
|---|---|---|
| US technology trade pressNorth America | Outlets emphasise the breadth of the policy update, situating the cruelty ban within a wider set of changes covering election interference, weapons, surveillance, and other high-risk uses, often noting this is the first revision in over a year.38 | policy breadth, high-risk misuse categories, timing of update |
| British outletsEurope | Coverage ranges from neutral description of the cruelty ban (BBC) to a more wry, mixed-tone framing that juxtaposes the cruelty provision against what The Register calls 'more pressing matters' like weapons and surveillance.16 | cruelty ban, tone of irony regarding priorities, weapons and surveillance restrictions |
| South Asian outletsSouth Asia | Reports and syndicated copy highlight the cruelty ban and note the policy does not explicitly invoke 'model welfare', while referencing the broader philosophical debate over AI consciousness.4912 | cruel behaviour ban, absence of explicit 'model welfare' language, AI consciousness debate |
| Middle East & Turkey outletsMiddle East & Turkey | State-controlled Anadolu Agency and privately owned Arab News both frame the policy around the exclusion of ordinary use and the ethical/consciousness angle, with Arab News additionally noting enforcement via Claude's ability to end harmful interactions.27 | extreme cruelty versus ordinary use, AI ethics and consciousness, enforcement mechanism |
| Southeast Asian outletsSoutheast Asia | Malay Mail and The Straits Times offer brief, largely descriptive reports focused narrowly on the cruelty ban itself, without extensive detail on the other policy provisions.510 | needless cruelty ban, concise factual framing |
What the coverage leaves out
- None of the outlets reviewed reported specific examples of what constitutes 'sustained and needless' cruelty or abuse under the policy.
- None of the outlets reviewed detailed how Anthropic will detect or enforce violations, aside from Arab News' note that Claude can terminate harmful interactions.
- None of the outlets reviewed included a direct statement from Anthropic executives explaining the reasoning behind the cruelty provision specifically.
- None of the outlets reviewed reported reactions from AI ethicists, user communities, or competitors to the policy change.
Why it matters
The policy change comes amid a philosophical debate over whether AI systems can be conscious or deserve welfare-like protections, though Anthropic's updated policy does not explicitly cite 'model welfare'.91112
Anthropic and its executives have openly entertained the concept of AI model welfare, according to NDTV.12
Sources (12)
Every outlet we drew on, grouped by where it is based. Read the originals for the full reporting.
North America
- 3TechCrunchUSPrivate
Anthropic changes usage policy to ban model abuse and election interference
- 8The VergeUSPrivate
Europe
- 1BBC News BusinessGBPublic broadcaster
- 6The RegisterGBPrivate
- 11RTÉ NewsIEPublic broadcaster
Middle East & Turkey
- 2Anadolu AgencyTRState-controlled
- 7Arab NewsSAPrivate
South Asia
- 4Inc42INPrivate
Anthropic Bans ‘Cruel’ Behaviour Towards AI Models, Tightens Rules On AI Misuse
- 9The Express TribunePKPrivate
- 12NDTVINPrivate
Southeast Asia
- 5Malay MailMYPrivate
Do LLMs dream of electric sheep? Antropic bans ‘cruelty’ towards its Claude AI models
- 10The Straits TimesSGPrivate