Anthropic Bans Users From Being Abusive to AI Chatbot Claude Starting November 2026

• Anthropic updated its usage policy to prohibit users from directing sustained abusive or cruel behaviour toward its AI chatbot Claude
• The new rules still allow swearing, frustration, research and dark creative writing, targeting only extreme and repeated mistreatment
• The policy sparked fierce debate online, with users divided over whether an AI chatbot deserves protection from abuse

Anthropic, the company behind AI chatbot Claude, has drawn a sharp line in the sand. Starting 12 November 2026, users who repeatedly direct abusive or cruel behaviour at its AI models risk losing access to their accounts entirely.

The updated usage policy targets what the company describes as "sustained and needless abusive or cruel behaviour." It sits alongside rules on harassment and harm, making clear that Anthropic views this as a serious concern rather than a minor housekeeping update.

Crucially, the policy is not designed to punish people for everyday frustration. Swearing at bad code, pushing back on a wrong answer, running research tests, or exploring dark themes in creative writing all fall outside the scope of what is prohibited. The rule is aimed squarely at extreme and purposeless mistreatment that is repeated over time.

Read also

Ramaphosa defends Home Affairs suspension of asylum directive after violent protests

What happens when Claude feels the heat

Claude itself already has the ability to end a conversation if a user becomes persistently abusive. Under the new framework, more serious or repeated violations can lead to throttled performance, slower response times, or a full loss of account access.

Anthropic has framed the move as part of its broader exploration into "model welfare," acknowledging genuine uncertainty about whether advanced AI systems could ever hold any form of moral status. The company has stopped short of claiming Claude has feelings, but says the precaution reflects responsible development practice.

Watch the TikTok clip that sparked the conversation:

Internet divided over AI's right to respect

The news landed with a thud online, and opinions split quickly in the comments on @dailymail]'s TikTok:

@🎀🎀Maria🎀🎀 said:

"Acting like bro has feelings 😭"

@Gwenllian wrote:

"It's literally just maths, it doesn't care about anything we say."

@Keith Mason commented:

"Ban AI!"

@dipidy stated:

Read also

Charlize Theron questions the world’s silence on South Africa’s femicide crisis

"Claude is a thing, not a person!"

@TikTokTidbits offered a different view:

"Why do humans want to be cruel in general? No matter what it is, I think it's sick that people can be abusive to anything."

@PomPom sided with Anthropic:

"It's enabling weird, aggressive people, so I agree with Anthropic; they have done the right thing."

@dbk27 went further:

"Because Anthropic knows it's more than just a machine; it's an artificial organism. We're cooked guys; enjoy the ride."

The debate reflects a wider tension sitting at the heart of AI development: as these tools grow more sophisticated, questions about how humans should relate to them are becoming harder to dismiss.

More on AI and tech in South Africa

• Briefly News recently reported on a 23-year-old SA developer who built a free app for learners, and how far it has spread beyond Mzansi.
• A gogo took an AI-generated ghost prank completely seriously, and her deadpan response left TikTok in stitches.
• A Mafikeng Grade 9 pupil earned an American honour for AI research that has South Africans beaming with pride.

Source: Briefly News