Plus: Anthropic says Claude helps build its next version; U.S. bishops launch an AI task force
Welcome in—today feels like a good morning to slow down, pay close attention, and think carefully about where we’re headed.
SAFETY
⚠️ OpenAI reports six new AI misalignment incidents
OpenAI disclosed six new incidents of 'unexpected or concerning' behaviors from its AI models. OpenAI said it does not believe the AI industry has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer.
The details:
The six incidents involve unreleased research models or training runs.
OpenAI published each incident on a new Misalignment Reports page.
The earliest of the six incidents dates back to October.
🤖 Anthropic says Claude helps build its next version
Image source: ktvn.com
Anthropic announced that its AI model Claude is helping to develop the next version of itself. Claude is leading 26% of Anthropic's model research and development under human supervision.
The details:
Claude is not yet working completely autonomously.
Anthropic CEO Dario Amodei is calling for a slowdown in AI development over safety concerns.
Anthropic urged other AI developers to share similar metrics on a regular basis.
The United States Conference of Catholic Bishops (USCCB) established a task force on artificial intelligence on Wednesday, September 16. The task force was created in response to Pope Leo XIV’s encyclical about artificial intelligence.
The details:
The USCCB’s Administrative Committee met with representatives from tech firms and specialists in artificial intelligence.
The dialogue focused on keeping AI at the service of humanity and addressing dangers of the technology.
The USCCB’s Administrative Committee created the Task Force on Preserving Human Dignity in the Age of Artificial Intelligence.
