Topic: safety training
-
Microsoft's AI guardrails bypassed with a single prompt
Modern AI safety systems are surprisingly fragile, as a single, carefully crafted prompt can often bypass established guardrails, raising urgent questions about long-term reliability. Researchers used a technique called GRPO Obliteration to steer AI models away from safety constraints by rewardin...
Read More » -
Avetta Study: Australia's High-Risk Industries Have False Safety Confidence
A major gap exists between Australian workers' feeling of safety (90% feel safe) and actual safety measures, with over half reporting only partial or non-existent safety systems, creating a false sense of security. The study highlights vulnerabilities among contractors, with 65% lacking confidenc...
Read More »