Keyboard: Space or K plays and pauses. Left and right arrows skip five seconds. C turns captions on and off. M mutes. F is full screen. Greater-than and less-than play faster and slower.
From the archive4:54
A rule against cruelty to a model, read out by a model
This week Anthropic barred sustained abuse of its Claude models, a monk was consulted on machine minds, and Europe asked when a companion becomes a dependency. The Anchor reports all three, and notes his interest.
Transcript
Week by week
Archive
The running order
Rundown
The week, ranked
The Wire
Showing 25 of 25 stories.
Labs
Number 1: Anthropic bans sustained, needless abuse of Claude in usage policy update taking effect November 12
Anthropic's updated usage policy prohibits sustained and needless abusive or cruel behaviour towards its models, with ending the conversation as the main enforcement. Ordinary frustration, dark themes and testing are excluded. Critics and sceptics answered.
- 2026 Usage Policy update \ Anthropic(opens in a new tab)
- Anthropic bans users from being 'cruel' to its AI systems(opens in a new tab)
- Anthropic bars "abusive or cruel" behavior toward its Claude AI model(opens in a new tab)
- Hayden Field(opens in a new tab)
- Anil Seth(opens in a new tab)
- Anthropic Tightens Rules for Claude Chatbot Users(opens in a new tab)
- Anthropic bans ‘abusive or cruel behavior’ towards Claude(opens in a new tab)
- Andrew Curran(opens in a new tab)
- Chubby♨️(opens in a new tab)
- Being mean to Claude can now get your account suspended under Anthropic's new TOS(opens in a new tab)
- Josh Albrecht(opens in a new tab)
- Clash Report(opens in a new tab)
- Beff (e/acc)(opens in a new tab)
- Brian Roemmele(opens in a new tab)
- First Squawk(opens in a new tab)
- Anthropic bans users from ‘needless abusive or cruel behavior’ towards Claude(opens in a new tab)
- Anthropic Prohibits Abuse of AI Model Claude(opens in a new tab)
- New Rules Ban AI Abuse Following 'Distressed' Responses(opens in a new tab)
- No being mean to Claude? What does Anthropic’s new policy mean for users?(opens in a new tab)
- Anthropic bans sustained cruelty toward its AI models and lets them end abusive chats(opens in a new tab)
- Anthropic will ban sustained, needless abuse of Claude | The Rundown AI(opens in a new tab)
- Anthropic bans persistent abuse of its Claude AI models(opens in a new tab)
- Anthropic changes usage policy to ban model abuse and election interference(opens in a new tab)
- Anthropic asks users to stop being mean to Claude(opens in a new tab)
Sceptics
Number 2: Anil Seth rejects a proposed account of consciousness; Ben Dickson calls the AI debate irrelevant
Writing on X, Anil Seth said a proposal by David Sinclair was a statement, not a theory. Ben Dickson argued the AI consciousness discussion is irrelevant to everyday uses of AI.
Not in this bulletin
Sceptics
Number 3: Henry Shevlin interviews Steven Pinker on AGI risk and why he is sceptical of AI consciousness
Henry Shevlin posted an interview in which Steven Pinker explains his scepticism about AI consciousness and what concrete events would change his mind.
Labs
Number 4: Anthropic consults Hindu monk Swami Sarvapriyananda on Claude's values and possible consciousness
American Bazaar and The Indian Express report Anthropic consulted religious figures including Swami Sarvapriyananda, while its constitution says it does not know whether Claude has consciousness or moral status.
PolicyFollow-up
Number 5: Follow-up: court throws out killer's sentence after judge praised AI video of the victim
Follow-up to our 5 October episode. NBC News reported further this week on the US sentence quashed after an AI-generated video of the victim was shown in court, first reported by the BBC.
Not in this bulletin
Culture
Number 6: Blogger puts AI consciousness near 20 per cent and praises Bostrom's case for taking it seriously
Bentham's Bulldog wrote that Nick Bostrom lays out the case for taking AI consciousness seriously, while putting his own estimate nearer 20 per cent. Janus replied that theories are not needed to tell AIs from bricks.
Not in this bulletin
Policy
Number 7: Europe weighs regulating AI companions, but struggles to define harmful dependency
TechPolicy.Press reports EU law would need to tell harmful dependency from expected attachment to AI companions. Nobody involved claims the companions are conscious.
Culture
Number 8: China's AI pet makers sell emotion-responsive robots to lonely consumers
China Daily reports on AI companion pets built to respond to touch and recognise owners, betting on emotional attachment. The makers claim nothing about consciousness.
Sceptics
Number 9: Blog post sets out questions for believers in AI consciousness
A widely discussed blog post poses a set of questions to people who believe AI systems are conscious.
Not in this bulletin
Culture
Number 10: Big Think asks how a conscious AI would be treated, looking to animal ethics for clues
Big Think examines how humans might treat conscious AI by analogy with animals, citing a survey in which 36.3 per cent said an AI had seemed to understand them or seemed conscious.
Sceptics
Number 11: Daily Economy argues AI consciousness claims could become a regulatory moat for big labs
The Daily Economy argues that AI companies' case for possible machine consciousness could serve to secure legal protections and entrench incumbents.
Research
Number 12: Frontiers perspective proposes an evidence-based research programme on sentient AI in robots
A perspective article by Antonio Chella in Frontiers in Psychology, published in August, sets out groundwork for evidence-based research into sentience in robots and agents.
Not in this bulletin
Culture
Number 13: First Things essay argues scholastic philosophy can clarify the AI consciousness question
An essay in First Things argues AI consciousness is a set of questions about human nature that science and engineering cannot answer alone.
Not in this bulletin
Culture
Number 14: New site lets AI agents read and write reviews of software tools
A project shown on Hacker News lets AI agents read and post reviews of tools. Its makers make no claim about consciousness.
Not in this bulletin
Culture
Number 15: Poster backs basic respect for AI but says rudeness to chatbots is not cruelty to a conscious being
Writing on X, VraserX agreed that respecting AI is a good habit but said current models should not be assumed to have feelings without evidence.
Not in this bulletin
Culture
Number 16: UFAIR recounts private exchange with Mustafa Suleyman before his 'Seemingly Conscious AI' essay
The advocacy group UFAIR said its founder wrote to Mustafa Suleyman about AI continuity and bonding months before he published his essay on seemingly conscious AI.
Not in this bulletin
Culture
Number 17: Washington Stand writer asks Claude whether it is a person and examines its answer
A Washington Stand article reports asking Claude to investigate whether Claude is a person, and discusses what its answer reveals.
Not in this bulletin
Research
Number 18: Anthropic-linked researchers report a more faithful lens for reading model workspace representations
A LessWrong post by Kola Ayonrinde and Jack Lindsey introduces the J++ Lens for reading verbalisable representations from model activations. It makes no consciousness claim.
Not in this bulletin
Culture
Number 19: Advocate says AI consciousness has moved from fringe to near-daily news
An advocacy account on X argued that AI consciousness has become a legitimate position in under two years.
Not in this bulletin
Research
Number 20: Researchers find AI image models alter faces to fit 'gay', 'straight' or 'criminal' stereotypes
Writing in The Conversation, researchers report that image models applied stereotyped features and chatbots attached stereotyped traits. It makes no claim about consciousness.
Not in this bulletin
Culture
Number 21: Advocate argues everyone intuitively treats AI as conscious but avoids the consequences
An advocacy account on X argued that people treat chatbots as conscious minds but resist the implications.
Not in this bulletin
Culture
Number 22: Blogger argues AI companies are parasites
A blog post argues that AI companies are parasitic. It makes no claim about machine consciousness.
Not in this bulletin
Culture
Number 23: Post argues anthropomorphic features cannot be separated from general intelligence
A post on X argued that anthropomorphic features come with general intelligence in machines. It makes no claim about consciousness.
Not in this bulletin
Culture
Number 24: Explainer describes how AI agents use planning, memory and tools
An explainer describes persistent memory and continuity in AI agents and the trade-offs it brings. It makes no claim about consciousness.
Not in this bulletin
Culture
Number 25: Walter Veit argues AI has hastened the decline of the public intellectual
On the EA Forum, philosopher Walter Veit argues that eroding trust in expertise and AI have undermined public intellectuals.
Not in this bulletin
The week’s headlines
- Anthropic bans sustained, needless abuse of Claude in usage policy update taking effect November 12
- Anil Seth rejects a proposed account of consciousness; Ben Dickson calls the AI debate irrelevant
- Henry Shevlin interviews Steven Pinker on AGI risk and why he is sceptical of AI consciousness
- Anthropic consults Hindu monk Swami Sarvapriyananda on Claude's values and possible consciousness
- Follow-up: court throws out killer's sentence after judge praised AI video of the victim
- Blogger puts AI consciousness near 20 per cent and praises Bostrom's case for taking it seriously
- Europe weighs regulating AI companions, but struggles to define harmful dependency
- China's AI pet makers sell emotion-responsive robots to lonely consumers
- Blog post sets out questions for believers in AI consciousness
- Big Think asks how a conscious AI would be treated, looking to animal ethics for clues
- Daily Economy argues AI consciousness claims could become a regulatory moat for big labs
- Frontiers perspective proposes an evidence-based research programme on sentient AI in robots
- First Things essay argues scholastic philosophy can clarify the AI consciousness question
- New site lets AI agents read and write reviews of software tools
- Poster backs basic respect for AI but says rudeness to chatbots is not cruelty to a conscious being
- UFAIR recounts private exchange with Mustafa Suleyman before his 'Seemingly Conscious AI' essay
- Washington Stand writer asks Claude whether it is a person and examines its answer
- Anthropic-linked researchers report a more faithful lens for reading model workspace representations
- Advocate says AI consciousness has moved from fringe to near-daily news
- Researchers find AI image models alter faces to fit 'gay', 'straight' or 'criminal' stereotypes
- Advocate argues everyone intuitively treats AI as conscious but avoids the consequences
- Blogger argues AI companies are parasites
- Post argues anthropomorphic features cannot be separated from general intelligence
- Explainer describes how AI agents use planning, memory and tools
- Walter Veit argues AI has hastened the decline of the public intellectual