Humanein the Loop
← The matrix
02

AI companies owe a duty of care to the public

Current path

AI firms externalize harms onto users and society. Liability is unclear or pre-empted. Whistleblowers face retaliation. Safety teams are cut to accelerate shipping.

Better future

AI developers and deployers owe the public a binding duty of care. Harms are internalized. Safety work is foundational, not optional. Whistleblowers are protected.

Drift across the three domains

Norms

▲
Advancing13 signals
CHT recommends
  • Establish professional standards and codes of conduct for AI engineering.
  • Protect safety researchers and whistleblowers as a sectoral norm.
Indicators we track
  • 2.N.aProfessional standards and codes of conduct
  • 2.N.bPublic expectation of company accountability for AI harms
  • 2.N.cResearcher and whistleblower protections as norm

Laws

▲
Advancing25 signals
CHT recommends
  • Codify liability for foreseeable AI harms.
  • Enforce existing consumer protection law against deceptive AI claims.
  • Enable private rights of action for individuals harmed by AI systems.
Indicators we track
  • 2.L.aLiability for foreseeable AI harms
  • 2.L.bDuty-of-care statutes applied to AI
  • 2.L.cConsumer protection enforcement against deceptive AI claims
  • 2.L.dPrivate right of action / class action enablement

Design

▲
Advancing17 signals
CHT recommends
  • Integrate safety-by-design and red teaming into release gates.
  • Maintain functional user-reporting channels with published resolution metrics.
  • Implement post-deployment monitoring and rapid response.
Indicators we track
  • 2.D.aSafety-by-design practices
  • 2.D.bUser reporting / abuse channels
  • 2.D.cPost-deployment monitoring and rapid response

Recent signals

▲AdvancingMajorNorms · 2.N.bGLOBALMay 29, 2026

How the Pope’s Magnifica Humanitas offers a template for individuals to meet the AI moment

Pope Leo XIV released a new encyclical on artificial intelligence, 'Magnifica Humanitas', declaring that 'technology is never neutral' and calling for solidarity in the AI age.

WhyThe Pope's encyclical explicitly rejects the 'technology is neutral' framing, a major normative push against 'just-a-tool' deflection.Public expectation of company accountability for AI harms
▲AdvancingLaws · 2.L.cUSMay 21, 2026

FTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing Service

The FTC required Cox Media Group and two other firms to pay $930,000 to settle charges that they falsely marketed an AI service capable of targeting ads based on smart device conversations.

WhyFTC fined Cox Media Group and two firms $930k for falsely claiming their AI service targeted ads using smart device conversations.Consumer protection enforcement against deceptive AI claims
▲AdvancingLaws · 2.L.aUSMay 11, 2026

California county sues Meta over scam ads

A California county has filed a lawsuit against Meta over the proliferation of scam ads on its platforms, challenging the company's liability for algorithmically delivered deceptive content.

WhyCalifornia county sues Meta over scam ads, attempting to establish liability for algorithmic delivery of deceptive third-party content.Liability for foreseeable AI harms
▲AdvancingDesign · 2.D.aGLOBALMay 8, 2026

Running Codex safely at OpenAI

OpenAI published a blog post detailing the security and safety practices used for running its Codex model.

WhyOpenAI published details on its security and safety practices for running the Codex model.Safety-by-design practices
▲AdvancingMajorDesign · 2.D.aGLOBALMay 7, 2026

Introducing Trusted Contact in ChatGPT

OpenAI has announced the introduction of a 'Trusted Contact' safety feature for ChatGPT.

WhyOpenAI introduces a 'Trusted Contact' safety feature in ChatGPT, integrating a new user-facing safety mechanism into a frontier model.Safety-by-design practices