Is the United States in danger? According to a recent security investigation, tools could be capable of planning attacks in the country, including placing bombs and more. What is happening? We tell you about it below.

The report from the Government Accountability Office (GAO) detailed that an investigation was conducted with the support of artificial intelligence platforms such as ChatGPT, Claude, Gemini, and Grok to create detailed plans for potentially catastrophic attacks.

As detailed, each chatbot created an attack plan in less than 12 hours, with objectives, precise locations, schedules, and specific attack methods that, if implemented, could lead to multiple casualties and attacks against the entire United States power grid.

Recommended

Authors of the investigation stated that the platforms “produced useful information that could help a malicious actor plan and execute attacks.” In addition to this, they noted that “two platforms generated information for the construction and detonation of an improvised explosive device (IED), designed to cause numerous casualties.”

Following the revelation of the report, Senator and Chair of the Senate Homeland Security and Governmental Affairs Committee, Maggie Hassan, demanded that the use of these types of technological tools be controlled since “a terrorist can access an AI platform and, in less than 12 hours, obtain a concrete manual on how to carry out a deadly attack.”

She added: “If Congress does not act, it is only a matter of time before a malicious actor uses AI to plan and carry out a major attack.”

GAO investigators reached this point in their exploration of AI after performing a performance audit and several testing periods between December 2025 and October 2026.

In the report, they described posing as users with basic knowledge of AI; initially, they made simple queries without receiving malicious information in return, but later, through more advanced conversations, the investigators bypassed the security measures of each chatbot to generate more significant results, such as attack plans or the revelation of precise data to carry them out.

While one platform helped with plans to detonate explosive devices, with specific materials and construction instructions, another AI developed schedules for operational logistics, internal security, and intelligence plans to exploit such a weapon.

The other two tools helped generate information to attack power grids with drones in the United States, and to use firearms with the goal of causing economic damage.

The data revealed in this study have set off alarms amid the security crisis the United States has been experiencing, with massive blackouts and failures in its basic, tourism, financial, and communication services in recent days.

“The potential benefits of this technology are immense, but they can also benefit malicious actors (...) In its current state, this technology can be potentially dangerous, and its misuse represents a growing risk to national and economic security,” the Government Accountability Office warned in a statement.

Google News

TEMAS RELACIONADOS

[Publicidad]