The Midas Project Watchtower
@safetychanges.bsky.social
We monitor AI safety policies from companies and governments for substantive changes.
Anonymous submissions: https://forms.gle/3RP2xu2tr8beYs5c8
Run by
@TheMidasProject.bsky.social
Anonymous submissions: https://forms.gle/3RP2xu2tr8beYs5c8
Run by
@TheMidasProject.bsky.social
Date: September 22, 2025
Company: Google
Change: Released v3 of their Frontier Safety Framework
Company: Google
Change: Released v3 of their Frontier Safety Framework
September 25, 2025 at 6:10 PM
Date: September 22, 2025
Company: Google
Change: Released v3 of their Frontier Safety Framework
Company: Google
Change: Released v3 of their Frontier Safety Framework
Date: Feb 26 - March 6, 2025
Company: Google
Change: Scrubbed mentions of diversity and equity from the mission description of their Responsible AI team.
Company: Google
Change: Scrubbed mentions of diversity and equity from the mission description of their Responsible AI team.
March 8, 2025 at 12:19 AM
Date: Feb 26 - March 6, 2025
Company: Google
Change: Scrubbed mentions of diversity and equity from the mission description of their Responsible AI team.
Company: Google
Change: Scrubbed mentions of diversity and equity from the mission description of their Responsible AI team.
Company: @anthropic.com
Date: February 27th, 2025
Change: Removed "White House's Voluntary Commitments for Safe, Secure, and Trustworthy AI," seemingly without a trace, from their webpage "Transparency Hub" (formerly "Tracking Voluntary Commitments")
Some thoughts in 🧵
Date: February 27th, 2025
Change: Removed "White House's Voluntary Commitments for Safe, Secure, and Trustworthy AI," seemingly without a trace, from their webpage "Transparency Hub" (formerly "Tracking Voluntary Commitments")
Some thoughts in 🧵
March 5, 2025 at 4:56 AM
Company: @anthropic.com
Date: February 27th, 2025
Change: Removed "White House's Voluntary Commitments for Safe, Secure, and Trustworthy AI," seemingly without a trace, from their webpage "Transparency Hub" (formerly "Tracking Voluntary Commitments")
Some thoughts in 🧵
Date: February 27th, 2025
Change: Removed "White House's Voluntary Commitments for Safe, Secure, and Trustworthy AI," seemingly without a trace, from their webpage "Transparency Hub" (formerly "Tracking Voluntary Commitments")
Some thoughts in 🧵
Over the past few weeks, a number of AI companies have released safety frameworks as promised at the Seoul AI Safety Summit.
Many others have not.
Most that do exist are weak or missing key components.
A 🧵 of the recently released policies and evaluation from
@themidasproject.bsky.social
Many others have not.
Most that do exist are weak or missing key components.
A 🧵 of the recently released policies and evaluation from
@themidasproject.bsky.social
February 14, 2025 at 11:17 PM
Over the past few weeks, a number of AI companies have released safety frameworks as promised at the Seoul AI Safety Summit.
Many others have not.
Most that do exist are weak or missing key components.
A 🧵 of the recently released policies and evaluation from
@themidasproject.bsky.social
Many others have not.
Most that do exist are weak or missing key components.
A 🧵 of the recently released policies and evaluation from
@themidasproject.bsky.social
Company: Google
Date: February 4, 2024
Change: Removed a ban on using AI technology for warfare and surveillance
URL: ai.google/responsibili...
Date: February 4, 2024
Change: Removed a ban on using AI technology for warfare and surveillance
URL: ai.google/responsibili...
February 4, 2025 at 11:12 PM
Company: Google
Date: February 4, 2024
Change: Removed a ban on using AI technology for warfare and surveillance
URL: ai.google/responsibili...
Date: February 4, 2024
Change: Removed a ban on using AI technology for warfare and surveillance
URL: ai.google/responsibili...
Company: Google
Date: February 4, 2024
Change: Released a new version of their Frontier Safety Framework
URL: deepmind.google/discover/blo...
A handful of changes in 🧵
Date: February 4, 2024
Change: Released a new version of their Frontier Safety Framework
URL: deepmind.google/discover/blo...
A handful of changes in 🧵
Updating the Frontier Safety Framework
Our next iteration of the FSF sets out stronger security protocols on the path to AGI
deepmind.google
February 4, 2025 at 11:10 PM
Company: Google
Date: February 4, 2024
Change: Released a new version of their Frontier Safety Framework
URL: deepmind.google/discover/blo...
A handful of changes in 🧵
Date: February 4, 2024
Change: Released a new version of their Frontier Safety Framework
URL: deepmind.google/discover/blo...
A handful of changes in 🧵
Company: Meta
Date: February 3, 2024
Change: Released a responsible scaling policy, called their Frontier AI Framework
URL: about.fb.com/news/2025/02...
Date: February 3, 2024
Change: Released a responsible scaling policy, called their Frontier AI Framework
URL: about.fb.com/news/2025/02...
Our Approach to Frontier AI | Meta
We’re sharing our Frontier AI Framework, which guides our consideration of risk in our model-release decisions, in line with the commitment we made at last year’s global AI Seoul Summit.
about.fb.com
February 4, 2025 at 11:09 PM
Company: Meta
Date: February 3, 2024
Change: Released a responsible scaling policy, called their Frontier AI Framework
URL: about.fb.com/news/2025/02...
Date: February 3, 2024
Change: Released a responsible scaling policy, called their Frontier AI Framework
URL: about.fb.com/news/2025/02...
Company: OpenAI
Date: January 13, 2024
Change: Adjusted the language on the o1 system card webpage, changing "o1" to "o1-preview."
URL: openai.com/index/openai...
Date: January 13, 2024
Change: Adjusted the language on the o1 system card webpage, changing "o1" to "o1-preview."
URL: openai.com/index/openai...
January 14, 2025 at 4:11 AM
Company: OpenAI
Date: January 13, 2024
Change: Adjusted the language on the o1 system card webpage, changing "o1" to "o1-preview."
URL: openai.com/index/openai...
Date: January 13, 2024
Change: Adjusted the language on the o1 system card webpage, changing "o1" to "o1-preview."
URL: openai.com/index/openai...
Company: Anthropic AI
Date: December 18, 2024
Change: Confusingly, within the last two days, Anthropic changed the "last updated" date on their Responsible Disclosure Policy to a date in June 2024, with no apparent substantive changes to the text of the policy.
Date: December 18, 2024
Change: Confusingly, within the last two days, Anthropic changed the "last updated" date on their Responsible Disclosure Policy to a date in June 2024, with no apparent substantive changes to the text of the policy.
December 19, 2024 at 12:52 AM
Company: Anthropic AI
Date: December 18, 2024
Change: Confusingly, within the last two days, Anthropic changed the "last updated" date on their Responsible Disclosure Policy to a date in June 2024, with no apparent substantive changes to the text of the policy.
Date: December 18, 2024
Change: Confusingly, within the last two days, Anthropic changed the "last updated" date on their Responsible Disclosure Policy to a date in June 2024, with no apparent substantive changes to the text of the policy.
Company: Cognition AI
Date: December 10, 2024
Change: Quietly updated terms of service concerning user data. Seemingly did not announce changes, nor have the changes been reflected in the "last updated" variable on their site.
URL: www.cognition.ai/pages/terms-...
Date: December 10, 2024
Change: Quietly updated terms of service concerning user data. Seemingly did not announce changes, nor have the changes been reflected in the "last updated" variable on their site.
URL: www.cognition.ai/pages/terms-...
December 11, 2024 at 6:40 AM
Company: Cognition AI
Date: December 10, 2024
Change: Quietly updated terms of service concerning user data. Seemingly did not announce changes, nor have the changes been reflected in the "last updated" variable on their site.
URL: www.cognition.ai/pages/terms-...
Date: December 10, 2024
Change: Quietly updated terms of service concerning user data. Seemingly did not announce changes, nor have the changes been reflected in the "last updated" variable on their site.
URL: www.cognition.ai/pages/terms-...
Company: Cohere
Date: November 21, 2024
Change: Released a complete rewrite of its usage policies. The spirit of the new document is similar — but includes a broader ban on “high-risk activities” and a stronger reporting commitment for detected CSAM.
URL: docs.cohere.com/docs/usage-p...
Date: November 21, 2024
Change: Released a complete rewrite of its usage policies. The spirit of the new document is similar — but includes a broader ban on “high-risk activities” and a stronger reporting commitment for detected CSAM.
URL: docs.cohere.com/docs/usage-p...
Usage Policy — Cohere
Developers must outline and get approval for their use case to access the Cohere API, understanding the models and limitations. They should refer to model cards for detailed information and document p...
docs.cohere.com
November 26, 2024 at 7:33 PM
Company: Cohere
Date: November 21, 2024
Change: Released a complete rewrite of its usage policies. The spirit of the new document is similar — but includes a broader ban on “high-risk activities” and a stronger reporting commitment for detected CSAM.
URL: docs.cohere.com/docs/usage-p...
Date: November 21, 2024
Change: Released a complete rewrite of its usage policies. The spirit of the new document is similar — but includes a broader ban on “high-risk activities” and a stronger reporting commitment for detected CSAM.
URL: docs.cohere.com/docs/usage-p...