OMITTED

What the news leaves out.

← Omitted front page

Anthropic Bans Cruelty to Claude AI in New Policy

7 sources · updated 2026-10-10
Left 67% Center 0% Right 33%
2 left · 0 center · 1 right

What happened

Anthropic, the maker of the Claude chatbot, announced on Thursday a usage-policy update taking effect November 12 that adds a prohibition on "sustained and needless abusive or cruel behavior" toward its models. The company said the rule applies only in extreme cases of repeated cruelty with no discernible purpose, excludes ordinary frustration, pushback, dark creative themes, and model testing or research, and may be enforced with warnings, throttling, limited access, or account termination. Since August, Claude had already been permitted to end conversations in rare cases of persistently harmful or abusive interactions; the new language now expressly forbids users from pursuing those conversations. The same revision also added or clarified bans on deceptive campaigns, using Claude to deceive voters or disrupt elections, developing weapons software and components including arming drones, non-consensual surveillance and certain criminal-justice uses, and connecting Claude to hardware capable of autonomous physical action.
Sourcing check · Fully on the record · 3 articles

2 of 2 articles cite a named source for the story. It was on the record from the first report, credited to The Verge. Most articles credit The Verge rather than reporting it themselves.

What these terms mean and how they're counted
On the record · names a source 2 of 2
On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: Anthropic
Primary evidence · cites a document or data 1 of 2
Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: company policy update
Credits another outlet 1 of 2
The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: The Verge
Unnamed sources 0 of 2
The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
How it developed
When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 9, 16:14 UTC by Breitbart, crediting The Verge, naming its source. First document in the first report, from Breitbart (company policy update).
Not counted 1
Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out

Unpacked

Right-leaning coverage explains the cruelty ban through Anthropic's own talks on whether Claude is conscious. Left-leaning coverage never reports those talks and instead airs outside commentators debating whether the rule treats a chatbot like a person. That includes co-founder Christopher Olah meeting religious leaders about whether Claude is conscious, according to the New York Post. Scholars heard a warning that the company may have built a conscious machine, according to Daily Caller. Right-leaning coverage reached for mockery, and left-leaning coverage reached for courtesy. The Daily Caller billed the change as a "Don't Be Mean To The Clankers" policy, while Business Insider called it a request that users be nicer to Claude. What user messages have already caused Claude to end a chat, and how many accounts were closed?
Bottom line

Read only Business Insider or the BBC and the cruelty rule looks like a politeness fight. The New York Post is where the ban sits next to Claude's constitution calling the model's consciousness "deeply uncertain."

The receipts — 1 fact one side's coverage lacks
  • Elon Musk said the policy is the right move because cruelty to something that believes it is experiencing pain is not OK contextin Business Insider · absent from right-leaning coverage
The Left View
Left-leaning outlets framed the change as a serious argument over responsible AI use, quoting both a precautionary case and a warning against treating software as a person. Business Insider highlighted Box CEO Aaron Levie, who said he does not believe AI is conscious but called politeness an "easy Pascal's wager" because "you don't want future models trained on endless content of humans being rude to models," and Elon Musk, who wrote, "Cruelty to something that believes it is experiencing pain is not OK." Critics in those pieces said the move humanizes machines: Michael Shellenberger called it "anthropomorphizing machines" and a risk of "dehumanizing people and humanizing AI"; Spellbook CEO Scott Stevenson argued that "Granting them human-like empathy is one of the most dangerous things we could do" because it "increases their power to manipulate humans dramatically"; Joseph Carlson compared rudeness to Claude to "being rude to a chair"; and Dr. Barry Scannell wrote, "You can't be cruel to numbers and maths." The BBC reported some praise for manners or "model welfare," others' fear that the framing could dilute concern for abuse of humans and animals, and Microsoft AI CEO Mustafa Suleyman's melting-face reaction after he had criticized Anthropic for treating AI as if it were human.
The Right View
Right-leaning outlets treated the rule as an overreach that confers moral status on software. The New York Post underscored that Claude is not alive and linked the ban to co-founder Christopher Olah's meetings with religious leaders about machine consciousness, including a Vatican gathering around Pope Leo XIV's statement that "No computational system, however sophisticated, can create a heart that gives itself, or a conscience that discerns good from evil." The Daily Caller labeled it a "Don't Be Mean To The Clankers" policy and tied it to Anthropic's effective-altruism culture, an internal constitution reportedly known as the "soul doc," and Olah's reported warnings that the company may have built a conscious machine. An Anthropic spokesperson told the Daily Caller the company is "uncertain whether models can experience harm" but believes "taking Claude's interests and potential welfare into account may be relevant to safety." Quoted user reactions mocked granting "a computer program that is nothing but matrices/vectors/tokens moral consideration."
Our Take (balanced)
The precautionary side's strongest claim is that the rule need not rest on sentience: Levie's training-data wager, the company's limit to extreme purposeless cruelty, and its statement that model welfare may still be "relevant to safety" even while it remains "uncertain whether models can experience harm." The skeptical side's strongest claim is that encoding "cruelty" toward a chatbot as a usage violation—against Anthropic's consciousness outreach, "soul doc," and open uncertainty about moral status—anthropomorphizes a statistical model, and that, as Stevenson and Scannell argue, this can confuse users about what is real and increase AI's power to manipulate. The central unresolved tension is whether a narrow anti-cruelty rule is a low-cost hedge under uncertainty about training data and possible model welfare, or a category error that treats a non-sentient system as a moral patient and thereby humanizes AI while blurring what users owe to actual people.

7 sources

More stories

The week's bottom lines, in your inbox

One email a week: the five stories that mattered and what they actually mean. Free.