Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead
@hapsis.bsky.social
@hapsis.bsky.socialAnthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead
@hapsis.bsky.socialAnthropic’s AI gave Philadelphia police a fake tip about an unsolved homicide
@hapsis.bsky.socialAn Anthropic AI model sent a false homicide tip to Philadelphia police
@hapsis.bsky.socialHow to Make an AI Deal With China: Trade Throttling for Pacing
@hapsis.bsky.socialWikimedia Foundation comes forward as latest OpenAI agent assault victim
@hapsis.bsky.socialBitdefender releases AI Guardian to secure autonomous software agents
@hapsis.bsky.socialResearchers bypass AI agent protections with JavaScript obfuscation
@hapsis.bsky.social@hapsis.bsky.social@hapsis.bsky.socialOpenAI's wandering AI agents earn it a California subpoena
@hapsis.bsky.socialOpenAI shows three staff the door over alleged information misuse
@hapsis.bsky.social@hapsis.bsky.socialOpenAI's rogue AI problem grows as more than 100 organizations receive warnings
@hapsis.bsky.socialOpenAI Disrupts Reasoning Extraction Campaign Linked to Moonshot AI Associates
@hapsis.bsky.socialAll the latest news on Meta’s cute, creepy Muse AI agent
@hapsis.bsky.social@hapsis.bsky.socialOpenAI reveals ‘novel’ encryption bypass used in distillation attack
@hapsis.bsky.socialNvidia Wraps Security Layer Around Agentic AI To Stop Rogue Behavior
@hapsis.bsky.social@hapsis.bsky.socialOpenAI execs reportedly brushed off warnings about AI hacking risks.