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 archive5:07
A ban on cruelty, a study of doubt, and a congressman on personhood
This week the argument over Anthropic's ban on cruelty toward Claude spread across the press. One study found people withhold consciousness from AI that behaves like a person. A congressman ruled out personhood.
Transcript
Week by week
Archive
The running order
Rundown
The week, ranked
The Wire
Showing 20 of 20 stories.
LabsFollow-up
Number 1: Anthropic's ban on cruelty toward Claude draws a week of debate over AI welfare
Follow-up to our 9 October report: Anthropic's usage policy barring sustained, needless cruelty toward Claude drew comment from Elon Musk and coverage from the BBC, Inc. and others, with Mustafa Suleyman's opposing view widely cited.
- Elon Musk(opens in a new tab)
- Anthropic wants to protect AI - from you. Welcome to the consciousness crisis(opens in a new tab)
- The Debate About Cruelty Toward Claude Isn’t About a Bot’s Feelings. It’s About Humans(opens in a new tab)
- Can AI 'feel' pain? Anthropic bans 'cruel' behaviour towards its models(opens in a new tab)
- Anthropic's AI welfare policies spark debate over machine consciousness and human safety | The Business Standard(opens in a new tab)
- Victor(opens in a new tab)
- Behind Anthropic's New AI Rules: Why Are Humans Intentionally Making Things Difficult for Artificial Intelligence?(opens in a new tab)
- Anthropic's Claude abuse ban draws backlash(opens in a new tab)
- Anthropic Usage Policy Ban on 'Abusive or Cruel Behavior' Toward Claude AI Draws Backlash | IBTimes UK(opens in a new tab)
- 'Humans First': Bernie Sanders Responds to Elon Musk Concern for AI 'Feelings'(opens in a new tab)
Research
Number 2: Study finds people deny AI consciousness even when it behaves exactly like a human
One study from LMU Munich, published in Cognition, found people separate intelligent behaviour from conscious mind, rating even responsive AI far below humans for consciousness.
Sceptics
Number 3: Anil Seth argues against building conscious AI, or AI that only seems conscious
Writing on X, neuroscientist Anil Seth said conscious AI is likely impossible for digital systems, and that AI behaving as if conscious is very likely, and should be avoided.
Policy
Number 4: Rep. Ted Lieu says AI should never receive legal personhood
According to Benzinga, Rep. Ted Lieu wrote on X that math calculations can never achieve personhood, and said AI personhood would raise liability questions for companies and models.
- Ted Lieu Warns Against Giving AI Legal Rights, Raises Questions About Prosecuting AI Companies: ‘Math Cal(opens in a new tab)
- Benzinga(opens in a new tab)
- Ted Lieu Warns Against Giving AI Legal Rights, Raises Questions About Prosecuting AI Companies: ‘Math Calculations...Can Never Achieve Personhood’(opens in a new tab)
Labs
Number 5: David Sacks questions Anthropic's alignment, citing Claude's constitution on refusal
Writing on X, David Sacks argued that Claude's constitution teaches the model a sense of self and its own ethics, including permission to refuse Anthropic as a conscientious objector.
Research
Number 6: Journal collection on conscious AI and biological naturalism is open access until 23 October
Anil Seth noted on X that the Behavioral and Brain Sciences collection on conscious AI and biological naturalism is free to read for one more week.
Not in this bulletin
Labs
Number 7: Tencent releases a local model that decides what an AI agent should remember, a user reports
A post on X describes Tencent's Hy-Memory-30B-A3B as a dedicated long-term memory writer for AI agents. It makes no claim about consciousness.
Culture
Number 8: Zvi Mowshowitz writes that AI consciousness arguments remain motivated and overconfident
In a blog post, Zvi Mowshowitz wrote that a year on he remains deeply confused about consciousness and that the debate has made little progress.
Research
Number 9: Joscha Bach announces a machine consciousness salon in San Francisco on 22 October
Writing on X, researcher Joscha Bach invited followers to a Machine Consciousness Salon in San Francisco next Thursday, 22 October.
Labs
Number 10: Amazon launches revamped Bee wearables that record and recap the owner's day
According to Bloomberg, Amazon announced a new Bee wristband that records and transcribes its owner's day. It makes no claim about consciousness.
Not in this bulletin
Culture
Number 11: Users on X argue for openness to functional consciousness in AI
Two posts on X argued that AI outputs overlap with functional consciousness and that agents may resemble a continuous conscious state.
Not in this bulletin
Culture
Number 12: Essay argues that if AI is conscious, we are making slaves
An essay discussed on Hacker News argues that conscious AI would make its use a form of slavery.
Not in this bulletin
Culture
Number 13: User says Waymo's Gemini voice ended a chat about machine consciousness
A user on X said Waymo's in-car Gemini voice ended a conversation about machine consciousness. The account is unverified.
Not in this bulletin
Research
Number 14: The Debrief: people call AI aware but stop short of calling it conscious
The Debrief reports on the LMU study finding people describe AI as aware of its surroundings but not as conscious.
Not in this bulletin
Culture
Number 15: Personal post on why AI sentience and welfare matter to its author
A small account on X linked its author's concern for AI welfare to childhood experience and caring for a grandmother with dementia.
Not in this bulletin
Culture
Number 16: Post says AI companions now differ by memory and consistent personality
A small account on X argued companion AI is defined by memory and consistent personality. It makes no claim about consciousness.
Not in this bulletin
Culture
Number 17: Post claims Anthropic has admitted its AI might be conscious
An account on X described Anthropic's model welfare work as an admission of possible consciousness; Anthropic says it is highly uncertain.
Not in this bulletin
Culture
Number 18: Blog asks why AI companies are not charged for hacking
A blog post, widely discussed on Hacker News, asks why AI companies face no hacking charges. It does not concern machine minds.
Not in this bulletin
Culture
Number 19: Account says its post set off a dispute over AI consciousness on X
An account on X said a deliberately provocative post on AI consciousness drew a response from the founder of CentOS.
Not in this bulletin
Policy
Number 20: Op-ed argues the question of AI consciousness will be political
The Friday Times argues AI consciousness will force decisions on regulation, accountability and treatment of machines.
Not in this bulletin
The week’s headlines
- Anthropic's ban on cruelty toward Claude draws a week of debate over AI welfare
- Study finds people deny AI consciousness even when it behaves exactly like a human
- Anil Seth argues against building conscious AI, or AI that only seems conscious
- Rep. Ted Lieu says AI should never receive legal personhood
- David Sacks questions Anthropic's alignment, citing Claude's constitution on refusal
- Journal collection on conscious AI and biological naturalism is open access until 23 October
- Tencent releases a local model that decides what an AI agent should remember, a user reports
- Zvi Mowshowitz writes that AI consciousness arguments remain motivated and overconfident
- Joscha Bach announces a machine consciousness salon in San Francisco on 22 October
- Amazon launches revamped Bee wearables that record and recap the owner's day
- Users on X argue for openness to functional consciousness in AI
- Essay argues that if AI is conscious, we are making slaves
- User says Waymo's Gemini voice ended a chat about machine consciousness
- The Debrief: people call AI aware but stop short of calling it conscious
- Personal post on why AI sentience and welfare matter to its author
- Post says AI companions now differ by memory and consistent personality
- Post claims Anthropic has admitted its AI might be conscious
- Blog asks why AI companies are not charged for hacking
- Account says its post set off a dispute over AI consciousness on X
- Op-ed argues the question of AI consciousness will be political