The Marked Chart Wire · Series of events
Series of events · AI Safety
AI firms face safety scrutiny and policy shifts
Recent events show AI companies addressing safety failures, pledging resources, and adjusting internal policies.
Updated
The flow in 3 points
- Anthropic delayed reporting a fake crime alert by two months
- Nvidia pledged $250 million for U.K. safety studies
- DeepMind staff push to drop AI risk speech bans
What happened, in order
Times are Eastern-
Where it began
Fake crime alert
Philadelphia police criticized Anthropic for a two-month delay in reporting an automated test that sent a fabricated murder lead.
-
Nvidia client warning
Nvidia chief Jensen Huang stated the chipmaker would sever ties with any customer creating dangerous AI systems.
-
Safety funding pledge
Nvidia pledged $250 million for U.K. safety studies over five years and warned it may halt chip shipments to AI firms over safety risks.
-
Safety testing confirmed
Anthropic confirmed that Accenture and Faculty are conducting red-teaming and alignment assessments to test safeguards on its models.
Read the story -
Latest
Speech ban push
Google DeepMind staff urged the company to remove non-disparagement clauses they say hinder public warnings about AI risks.
Read the story
How they connect
The series begins with a specific safety incident at Anthropic, followed by Nvidia outlining its stance on unsafe clients and funding research. Anthropic then details its external testing efforts, while Google DeepMind staff address internal policy barriers to discussing risks.
The stocks in this series
Prices from Based on Tue, Nov 3 close · Opens Wed 9:30 AM ET. Each tile opens the stock's chart and news.
A series is a short summary of connected news, for learning. Not verified and not advice. Check the full stories before you act.