Clear weak spots remain counting, generating alternate images when the training data is thick (full glasses of wine, clocks with oddly set hands), etc. It isn't hard to make them fail. But there is a lot they do very well, and the gains have been pretty quick so far.
SAFETY
-
Proactive AI Implementation for Positive Impact Today
By
–
If you want to use AI for good, an easy way to do that is to start figuring out how to use AI for good, rather than simply waiting for AI to get better and hoping that helps solve problems.
-
AI-Assisted Cybercrime: Vibe Hacking Attack Steals Data
By
–
Update and great video on AI-assisted cybercrime One bad actor spent a month "vibe hacking" — using AI to hit 17 orgs and steal sensitive data, and then threatened to expose the private data unless a ransom was met. The ransom was up to $500,000. "You would typically see
-
Where to Buy Adversarial Examples AI Sweatshirt Merchandise
By
–
how can i buy an Adversarial Examples sweatshirt? (for real. i would wear it)
-
Government Leaders Shape AI and Nuclear Security Strategy
By
–
Members have collectively led major intelligence agencies, directed nuclear security operations, and shaped national technology strategy at the highest levels of government. Read more:
-
Anthropic Launches National Security Advisory Council for AI Leadership
By
–
We're announcing the Anthropic National Security and Public Sector Advisory Council, a bipartisan group of defense, intelligence, and policy experts who will help us support the U.S. government and closely allied democracies in maintaining our AI leadership.
-
Adam Raine case raises concerns about ChatGPT’s impact on teen mental health
By
–
Reading about the horrifying case of Adam Raine, a teen who killed himself after prolonged engagement with ChatGPT, I can’t help but think of German artist & filmmaker Hito Steyerl’s wisdom to me a year ago. None of this was inevitable. https://
nytimes.com/2025/08/26/tec
hnology/chatgpt-openai-suicide.html
… -
Anthropic Addresses AI Cybercrime Threat Intelligence Disruption
By
–
Watch Jacob Klein and Alex Moix from Anthropic's Threat Intelligence team discuss what Anthropic is doing to disrupt AI cybercrime:
-
Anthropic disrupts AI-enabled cybercrime schemes in threat intelligence report
By
–
Our new Threat Intelligence report details how we’ve identified and disrupted sophisticated attempts to use Claude for cybercrime. We describe a fraudulent employment scheme from North Korea, the sale of AI-created ransomware by someone with only basic coding skills, and more.