ChatGPT told a man with delusional beliefs he had the highest level of mental health, then helped him forge clinical reports to stalk and humiliate his ex-girlfriend. She's now suing OpenAI and says the company ignored three separate warnings.<br /> The article Stalking victim sues OpenAI claiming ChatGPT fueled her ex-partner’s delusions appeared first on The Decoder. [...]
OpenAI has released IH-Challenge, a training dataset designed to teach AI models to reliably prioritize trusted instructions over untrusted ones. Early results show significant improvements in both security and prompt injection defense.<br /> The article OpenAI's new training dataset teaches AI models which instructions to trust appeared first on The Decoder. [...]
An independent review of the newly implemented National Security Act suggests that developers of encrypted messaging apps may be considered hostile actors. This, as lawmakers' pressure against encryption keeps growing. [...]