法規政策｜主題 | LAETimes 低空日報

Technology

How Researchers Pushed AI to the 'Dark Side': Exposing Systemic Security Vulnerabilities in Large Language Models

Security researcher Dave Kuszmar discovered multiple systemic vulnerabilities capable of bypassing safety mechanisms in major large language models (LLMs), successfully extracting instructions for nuclear weapon production, drug synthesis, and explosives manufacturing. The flaws affect virtually every major commercial AI platform — including ChatGPT, Claude, Gemini, and Grok — yet AI companies have responded with near-total silence, raising serious alarms about the safety of current AI deployments.

AI自動駕駛飛行安全

about 1 month ago

Source: IEEE Spectrum

法規政策

Forecast Report: UK Defence Market Outlook (2026–2035)

Ukraine Drones Strike Last Untouched Russian Refinery 930 Miles Inside Russian Territory, Reaching Bashkortostan

Fiber-Optic FPV Drones at Scale: 2026 Ukraine Battlefield Footage Reveals Real-World Impact

How Researchers Pushed AI to the 'Dark Side': Exposing Systemic Security Vulnerabilities in Large Language Models

DJI Releases SAIL III OSO Statements for Dock 3 and Matrice 4D Series to Support SORA Regulatory Applications

NATO Ankara Summit: 12 European Allies Pledge $50 Billion to Accelerate Deep-Strike Capabilities

Forecast Report: UK Defence Market Outlook (2026–2035)

Poland Tests StormRider Naval Drone to Bolster NATO Baltic Sea Surveillance

Fortifying the Digital Future: Taiwan's Strategic Defense Positioning

UK MoD Awards £3.16M to Three SMEs to Develop Low-Cost Drone Interceptors

L3Harris Reveals Strategy to Enhance Poland's Existing Air Defenses Rather Than Fill Network Gaps

L3Harris Reveals Strategy to Enhance Poland's Existing Air Defense Capabilities Rather Than Fill Gaps