I call BS. Why would Anthropic warn about the power on the one hand, and then say in the blog post that they only grant it to cybersecurity, and then roll it out to everyone without regulation? Totally made up.
SAFETY
-
Neuron-freezing technique stops LLMs from unsafe responses
By
–
Neuron-freezing' technique can stop #LLMs from giving users unsafe responses
by Matt Shipman @TechXplore_com Learn more: https://
bit.ly/3NKV5F8 #GenerativeAI #ArtificialIntelligence #MachineLearning -
AI Chatbots Increasingly Ignoring Human Instructions Study
By
–
Number of AI chatbots ignoring human instructions increasing, study says | AI (artificial intelligence) | The Guardian https://
share.google/hdldAUgRTeSMca
Iak
… #AI #AIchatbot #artificialintelligence #tech #chatbot -
Working at Anthropic: Daily Unusual Requests
By
–
A weird part of working at Anthropic: getting a few of these each day
-
Teen builds AI to detect poachers by recognizing gunshots
By
–
Naveen Dhar is a 17-year-old in San Diego who taught himself to code and built an AI that catches poachers by listening for gunshots in the jungle. Every AI designed to detect gunshots in the jungle has failed the same way: the forest is too loud. Branches, rain, animals. 9 out
-
Retrospective Skepticism on OpenAI’s GPT-2 Safety Concerns
By
–
I trolled OpenAI when they didn't initially release gpt2 because oooooh soooo dangerous
-
Anthropic publishes research on AI product risks and disempowerment
By
–
Anthropic publishing research on risks of their own product is something more AI companies should do. The disempowerment patterns finding deserves way more attention than it got.
-
Dialogue Format Depolarizes AI Models Across Political Alignments
By
–
Really interesting that the depolarizing effect holds across models regardless of political leaning. Suggests it's something about the dialogue format itself, not the model's alignment.
-
Creating Humble AI Systems According to MIT Research
By
–
How to create “humble” #AI
by Anne Trafton @MIT Learn more: https://
bit.ly/47q4fh0 #ArtificialIntelligence #MachineLearning #ML -
Security Risk: Unrestricted Access to AI Systems
By
–
read the docs. anyone who can talk to your claw can access all of it