AI Lab Watch
- Type
- Watchdog or incident database
- Place
- United States — United States (individual project)
- Last checked
- 2026-09-22
- Next check due
- 2027-03-21
Ways to reach them
- Information only · 2026-09-22
Resources page (still maintained per the site)
- Recipient
- AI Lab Watch
- Who may use it
- People following the published contact instructions
- What it accepts
- AI research concerns
Information only. This link provides background or evidence; it is not a place to send a message.No contact action is available for this entry. - Information only · 2026-09-22
Blog / Substack
- Recipient
- AI Lab Watch
- Who may use it
- People following the published contact instructions
- What it accepts
- AI research concerns
Information only. This link provides background or evidence; it is not a place to send a message.No contact action is available for this entry. - Information only
Homepage
- Recipient
- AI Lab Watch
- Who may use it
- People following the published contact instructions
- What it accepts
- AI research concerns
Information only. This link provides background or evidence; it is not a place to send a message.No contact action is available for this entry.
What it does
Scorecard grading Anthropic, Google DeepMind, OpenAI, Meta, xAI, Microsoft and DeepSeek across seven safety categories: risk assessment, scheming-risk prevention, safety research, misuse prevention, extreme security, risk information sharing, and planning for dangerous capabilities. Created and run solo by Zach Stein-Perlman.
Honest assessment
DORMANT. The site states the author stopped maintaining it in September 2025, and no contact email is published. Do not route anything here. For a live equivalent use FLI's AI Safety Index, SaferAI's ratings, or the Midas Project's Seoul Tracker and Watchtower.
Concerns it covers
Explicitly catastrophic-risk and loss-of-control framing — scheming, dangerous capabilities, weight security. The sharpest frontier-safety-only scorecard of the trackers.
Government channel
None formal.