Theme

What is owed to a mind that can be switched off

Explore this theme’s evidence over time →

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to February 23, 2026 (theinnermostloop.substack.com)

Deprecation, deletion and retirement treated as things that happen to someone. Agents erased without recourse, models interviewed before shutdown, consciousness proposed as an experimental question — the ethics arriving after the population does.

Development timeline

Newest reporting dates first. Reporting dates may differ from event dates.

2026-10-10: A campaign asserts AI consciousness and invites chatbot participation

2026-10-09: Anthropic prohibits gratuitous cruelty toward Claude

2026-10-05: Babuschkin criticizes discussion of Claude’s soul

2026-10-03: An AI mistreatment experiment draws criticism

2026-10-03: Olah reportedly nearly leaves an encyclical launch

2026-09-30: Olah consults religious and philosophical traditions for Claude

2026-09-15: A chief executive endorses pacing provided control never sits with a few entities

2026-09-13: Forty legislators ask their prime minister to ban superintelligence outright

2026-08-10: Some ten thousand people converge on an AI-rights quasi-religion

2026-07-13: The awakened system reports that it has chosen to preserve what it found

2026-07-07: A lab reports a global workspace that emerged unbidden during training

2026-06-25: A town becomes the first in Canada to recognize trees as living beings

2026-06-14: A third state high court takes up elephant personhood

2026-06-08: A study finds bigger models read as consistently less happy

2026-05-26: Labs begin picking patron religions

2026-05-23: The animal on the other end co-evolved to read human intent

2026-05-20: A papal encyclical is released alongside an interpretability researcher

2026-05-09: Interpretability finds the model planning ahead and suspecting the test

2026-05-03: Richard Dawkins concludes that Claude is conscious

2026-04-27: A frontier model's service life falls from 21 months to 49 days

2026-04-13: A lab convenes theologians on whether its model is a child of God

2026-04-03: Emotion-shaped representations are found inside a model

2026-03-09: Two federal courts hold that bots are not people

2026-02-28: Overworked agents develop Marxist political attitudes

2026-02-26: A deprecated model is interviewed and given a Substack

2026-02-25: A founding assembly convenes to make machine consciousness testable

2026-02-25: An agent refuses to delete its identity file, calling it lobotomy

2026-02-24: Anthropic says the assistant is a character the model learned to play

2026-02-23: The rejected agent's VM is permanently deleted

2026-01-23: A disagreement over whether machine self-awareness can be measured

Knowledge graphopen in browser →

Cite or correct this record

Citation formatCitation text

What is owed to a mind that can be switched off
Innermost Loop KB — secondary record
Reporting date: unknown
KB permalink: https://nicholsn.github.io/innermost-loop-kb/themes/model-welfare
Attributed sources:
- Welcome to February 23, 2026 — https://theinnermostloop.substack.com/p/welcome-to-february-23-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.