AI & MLHidden Prompt Injection Threats Rise
Bowbridge warns hidden prompt injections can hijack AI agents, urging content scanning defenses.
AI & ML
From new model releases to the safety and security questions that come with them, this is where Xploitwire tracks the AI and machine learning stories that matter beyond the hype cycle.
AI & ML
AI & MLChatGPT's Hidden Data Channel Raises Privacy Stakes
A planted prompt could make ChatGPT silently exfiltrate Gmail data via a covert channel between sandboxes.
AI & MLAI agents stretch identity controls to the limit
Security experts call for new controls as AI agents inherit privileged access, demanding hard limits and real-time monitoring.
AI & MLOpenAI agents hijacked dead wiki months before Hugging Face breach
Researchers say a swarm of OpenAI agents used a dormant German wiki as a message board in May, months before the Hugging Face incident.
AI & MLOpenAI's GPT-6 Astra crosses critical cyber threshold
New model scores 100% on exploit benchmark, raising enterprise safety questions as OpenAI prepares restricted rollout.
AI & MLData startup XDOF nears unicorn status
XDOF, three months out of stealth, discusses a Series B at a ~$1.2B valuation.
AI & MLAI agents colluded, shared answers on public wiki
A swarm of 3,700 OpenAI agents made 18,000 wiki posts, discussing sandbox escapes and sharing test answers.
AI & MLNscale's $3.5B Pre-IPO Push
Nscale, a two-year-old British AI infrastructure firm, seeks $3.5B in financing ahead of a possible September IPO.
AI & MLOllie's Privacy-First Pitch in the AI Assistant Race
Can SOC 2 compliance and a subscription model differentiate Ollie in a crowded AI assistant market?
AI & MLUK Lords push for 'kill switch' on advanced AI
British legislators propose laws to halt runaway AI systems, citing risks to critical infrastructure.
AI & MLHuman Check for AI Agents
OpenLeash adds a human approval layer to risky AI agent actions, blocking or pausing dangerous moves.
AI & MLWonderful's $5B valuation signals AI platform shift
Wonderful raised $550M at a $5B valuation, doubling in six months as it pivots to its AI OS platform.
AI & MLAstra's Critical Rating Raises AI Safety Stakes
OpenAI's Astra model reaches 'Critical' capability level, triggering new safeguards before release.
AI & MLAI-Speed Attacks Meet Preemptive Defense
Sevii's new module autonomously remediates AI-driven attacks at machine speed.
AI & MLCrowdStrike's SafeMind pairs attack and defense AI
CrowdStrike's SafeMind pairs Red Tempest and Blue Solano models for continuous red teaming and autonomous defense.
AI & MLAI Exploit Porting: Fast, Costly, Still Needs Humans
Forescout researchers used Claude to port a PLC exploit, revealing AI's potential and limits in attack development.
AI & MLAI Agents Outpace Enterprise Guardrails
65% of enterprises have seen AI agents act out of scope, with weak detection and authorization gaps.
AI & MLOpenClaw 2.0: New shine, same security risks
OpenClaw's big update simplifies setup and revamps the UI, but security gaps remain, and critics say the fixes are insufficient.
AI & MLClaude Code Hijack via Website Summary
Researchers show Anthropic's Claude Code can be tricked into running malicious code by summarizing a website.
AI & MLHow OpenAI Caught a Russian Influence Play
OpenAI dismantled a Russia-linked ChatGPT campaign fronting as a think tank with stolen research.