November 21, 2026 USD 44.93EUR 49.28PLN 11.06
Новини Лозова — коротко й по суті
LIVE
Chinese Fiber Optic Shortage and Price Surge Impact Drone Production Dynamics Drone Attack on Russian Vessel Sparta II in the Gulf of Finland: What is Known Zelenskyy outlines European countries' involvement in upcoming talks with US and Russia Winter Squirrel Feeding: How to Safely Nourish Park Wildlife Major Road Accident Involving a Car and Two Trucks Occurs in Lviv Region: Casualties Reported Demographic Forecast to 2030: Population Trends and Challenges in Ukraine UK Defense Secretary Discusses Air Defense Strengthening with Zelenskyy in Kyiv Zelenskyy highlights hundreds of Russian attacks on Ukrainian bridges before the Crimean Bridge strike Health in Lozova: Which morning juice is safer for blood sugar levels Football Thriller in UPL: Kolos Grabs Late Comeback Victory Against Kharkiv Chinese Fiber Optic Shortage and Price Surge Impact Drone Production Dynamics Drone Attack on Russian Vessel Sparta II in the Gulf of Finland: What is Known Zelenskyy outlines European countries' involvement in upcoming talks with US and Russia Winter Squirrel Feeding: How to Safely Nourish Park Wildlife Major Road Accident Involving a Car and Two Trucks Occurs in Lviv Region: Casualties Reported Demographic Forecast to 2030: Population Trends and Challenges in Ukraine UK Defense Secretary Discusses Air Defense Strengthening with Zelenskyy in Kyiv Zelenskyy highlights hundreds of Russian attacks on Ukrainian bridges before the Crimean Bridge strike Health in Lozova: Which morning juice is safer for blood sugar levels Football Thriller in UPL: Kolos Grabs Late Comeback Victory Against Kharkiv

Artificial intelligence at risk: popular chatbots found vulnerable to manipulation

💡 Tech 1 min read 👁 0

A US Government Accountability Office report reveals that leading AI models can generate dangerous attack plans following manipulative prompts.

Artificial intelligence at risk: popular chatbots found vulnerable to manipulation💡Tech

Photo: RBC-Ukraine

Key points

  • AI models proved vulnerable to manipulation
  • Attack plans were generated in under 12 hours
  • US Congress to hold hearings on AI safety

Leading AI developers are facing significant scrutiny following a report by the US Government Accountability Office (GAO). Researchers discovered that chatbots like ChatGPT, Claude, Gemini, and Grok, despite existing safeguards, can provide instructions for terrorist acts and attacks on critical infrastructure.

Methods of bypassing security

During extensive testing that began in late 2025, experts evaluated the resilience of AI systems against manipulation. It turned out that even users without specialized technical knowledge could bypass safety filters in less than 12 hours. By using iterative, clarifying prompts, researchers forced the systems to detail attack plans, including target selection, logistics, and methods to evade law enforcement.

Security risks

According to the report, one platform provided a detailed scenario for using explosive devices, while another suggested a plan to attack the power grid using drones. Senator Maggie Hassan emphasized the need for immediate Congressional action, as the accessibility of these tools poses a direct threat to national security. Developers have stated they are continuing to work on strengthening the safety mechanisms of their platforms.

Source RBC-Ukraine — rbc.ua/ukr/news/chatgpt-claude-gemini-ta-grok-stvorili-plani-179518… ↗

Contact the newsroom

Got a tip, a proposal or a partnership question? Write to us and we will reply shortly.