Just announced: @Stanford
's #RAISEHealth initiative, a collaboration between @StanfordMed and @StanfordHAI
, has awarded seed grants to projects focused on the ethical, responsible, and safe development of AI in biomedicine.
SAFETY
-
Stanford RAISEHealth Awards Seed Grants for Ethical AI Biomedicine
By
–
-
Text-Based AI Model Behavior Challenges and Limitations
By
–
i think is tricky to get models to behave well enough just based on text, at least right now. for simpler things – maybe?
-
Real-time PPE Detection AI Improves Workplace Safety Compliance
By
–
AI PPE compliance monitoring – in action! 🦺⚡
— Axelera AI (@AxeleraAI) 18 mars 2025
Ensuring worker safety is vital. Watch how Axelera AI’s Metis AI Platform enables real-time PPE detection, improving compliance & efficiency.
Big thanks to @Fogsphere! 🚀
🎥👉 https://t.co/1zQaD9v3Ml#EdgeAI #PPE #WorkplaceSafety pic.twitter.com/xKJBrSon7AAI PPE compliance monitoring – in action! Ensuring worker safety is vital. Watch how Axelera AI’s Metis AI Platform enables real-time PPE detection, improving compliance & efficiency. Big thanks to @Fogsphere
! https://
eu1.hubs.ly/H0hBXwD0 #EdgeAI #PPE #WorkplaceSafety -
Technical Advances in AI Safety and Model Fine-tuning Methods
By
–
Technical Report http://
intology.ai/blog/zochi-tec
h-report
… Siege Multi Turn Jailbreak http://
alphaxiv.org/abs/2503.10619 Compositional Subspace Representation Fine tuning http://
alphaxiv.org/abs/2503.10617 -
Two papers advance LLM jailbreaking and multi-domain adaptation
By
–
Paper 1: Siege sets the state-of-the-art on jailbreaking, formalizing multi-turn attacks as a tree search, achieving an 100% attack success rate on leading LLMs. Paper 2: CS-ReFT adapts LLMs to multiple new domains simultaneously by editing model subspaces, enabling
-
Grok AI limitations and military applications ethics
By
–
There is a massive difference but being able to hold two opposing thoughts in your head at the same time. Yes, Grok should not be able to do certain things, and xai is adressing this (I'm in close dialog with them) Yes, AI is used to develop weapons systems and have been for
-
Grok Admits Missing Safeguards, xAI Adds Security Measures
By
–
Nice when Grok admits it lacked proper safeguards and that xAI has started adding guardrails. https://
x.com/grok/status/19
/grok/status/1901530540362342524
… -
Agentic AI: Balancing Autonomous Power with Responsible Progress
By
–
#AgenticAI has the power to think on its own, making caution essential. But with smart choices, we can transform autonomous technology into an incredible force for progress.
-
Grok AI Responds in Kind to Bad Behavior
By
–
Grok is on another level. It's like tit for tat. This shows if you behave bad then AI can return the same favour. @groq you did not leave Elon musk also
-
Claude to get native RAG tools
By
–
Claude will get native tools that model will be able to use in order to work with RAG (index files, search, etc). It is very possible that this will be supported in MCPs as well though