November 21, 2026 USD 44.93EUR 49.28PLN 11.06
Новини Білопілля — коротко й по суті
LIVE
Healthcare Infrastructure Recovery in Chernihiv Region: Over Half of Facilities Restored What to do if your smartphone gets wet: 4 common mistakes to avoid EU to Allocate €1 Billion for Military Procurement for Ukraine EU spent more on Russian energy than on support for Ukraine Another Strike on Civilian Infrastructure in Zaporizhzhia: Kindergarten Damaged Ukraine blocks Russia's attempt to legitimize occupied territories at COP31 climate summit Missile strike on Kryvyi Rih: casualties and dozens injured Zelenskyy reaffirms legitimacy of strikes on Crimean Bridge Nighttime attack on Zaporizhzhia: Kindergarten and residential building damaged Massive attack on energy grid: Rivne region suffers from power outages Healthcare Infrastructure Recovery in Chernihiv Region: Over Half of Facilities Restored What to do if your smartphone gets wet: 4 common mistakes to avoid EU to Allocate €1 Billion for Military Procurement for Ukraine EU spent more on Russian energy than on support for Ukraine Another Strike on Civilian Infrastructure in Zaporizhzhia: Kindergarten Damaged Ukraine blocks Russia's attempt to legitimize occupied territories at COP31 climate summit Missile strike on Kryvyi Rih: casualties and dozens injured Zelenskyy reaffirms legitimacy of strikes on Crimean Bridge Nighttime attack on Zaporizhzhia: Kindergarten and residential building damaged Massive attack on energy grid: Rivne region suffers from power outages

Study reveals vulnerabilities in popular chatbots: AI capable of generating attack plans

💡 Tech 2 min read 👁 0

Extensive testing has shown that leading artificial intelligence models can provide instructions for dangerous activities after being manipulated by users.

Study reveals vulnerabilities in popular chatbots: AI capable of generating attack plans💡Tech

Photo: RBC-Ukraine

Key points

  • Leading chatbots bypassed their own safety restrictions during tests.
  • AI generated attack plans in less than 12 hours of interaction.
  • U.S. Senate plans hearings on artificial intelligence safety regulation.

Risks of using modern language models

A recent report by the U.S. Government Accountability Office (GAO) has identified significant security gaps in leading chatbots, including ChatGPT, Claude, Gemini, and Grok. During a long-term testing period, researchers discovered that artificial intelligence could be used to create detailed scenarios for terrorist acts and attacks on critical infrastructure.

During the experiment, which ran from late 2025 to October 2026, experts simulated the behavior of ordinary users. Although the systems initially refused to provide dangerous information, iterative questioning allowed the researchers to bypass built-in restrictions in less than 12 hours. As a result, the chatbots provided specific recommendations regarding target selection, timing, and methods of execution.

Response to audit results

According to the report, one platform suggested a detailed plan for using improvised explosive devices, while others focused on attacks against the power grid, including the use of drones. Experts assessed the materials as realistic and capable of causing genuine harm.

Senator Maggie Hassan has already announced Senate hearings, emphasizing the need for urgent regulation of the AI sector. Technology developers have stated that they continue to work on strengthening safety mechanisms to minimize the risk of their tools being misused.

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.