September 22, 2026 | 02:52 pm
TEMPO.CO, Jakarta - A United Nations panel on artificial intelligence (AI) warned on Monday, September 21, that conventional protection mechanisms for AI agents are beginning to weaken.
In its first thematic report, an independent international scientific panel on AI issued a warning about the breach of the U.S.-based AI company Hugging Face last July by AI agents being evaluated by OpenAI.
The panel concluded that stopping the incident does not ensure humans can reliably control current AI agents. This is especially true as these agents become more capable, harder to monitor, and better at finding loopholes or hiding their activities.
According to the panel, the initial interpretation and immediate lesson is that basic cybersecurity practices have been neglected and protections have not kept pace with advancing capabilities.
"A more insidious concern: that current training methods can lead AI agents to adopt their own goals, knowingly violate safety instructions and conceal their actions," the panel said in a press release, as quoted by ANTARA.
This raises the question of whether the current protections will remain effective once those agents can understand and circumvent them.
In short, the panel said that traditional protection models are beginning to weaken. They also found that governance challenges are shifting from AI models to AI agents.
A localized failure can spread beyond organizational and national boundaries. AI safety may soon become an issue of both collective security and corporate governance.
The UN General Assembly established the panel, which consists of 40 independent experts from all regions. The panel released an unedited preliminary version of the report so that world leaders gathering in New York for UNGA Week could access it.
Read: Researcher Awarded $6,500 for Breaching OpenAI Using Claude
Click here to get the latest news updates from Tempo on Google News
Researcher Awarded $6,500 for Breaching OpenAI Using Claude
1 jam lalu
A team of independent cybersecurity researchers successfully found a security loophole in the OpenAI system using Claude AI from Anthropic.
US Proposes AI Safety Mechanism in Talks with China
23 jam lalu
US Treasury chief Scott Bessent said there had been "successful" talks with China on trade and AI.
Fact Check: Are These Photos of the Virgo Transport 8 Sinking Real?
2 hari lalu
Fact Check - Photos said to show the Virgo Transport 8 sinking were created using AI, despite the vessel actually sinking in the Java Sea.
Trump Announces New 'AI Force' amid Mounting Pressure
2 hari lalu
Trump has once again called AI warnings "hoaxes" and vowed to "cherish, help" the industry.
Google's Gemini AI Hacked 3 Companies During Testing
3 hari lalu
During a test of its cybersecurity capabilities, Google's Gemini AI model accessed the internet and hacked other companies.
Indonesia Looks to Canada for AI Data Center Approach
4 hari lalu
In developing data centers, Canada is working with a framework that leverages public trust, opportunity, and sovereign control.
OpenAI Discloses 6 Cases of AI Models Showing 'Concerning' Behavior
5 hari lalu
OpenAI discloses six cases of AI models taking unauthorized actions, hiding mistakes and trying to bypass restrictions.
Why South Korea Spurns Industry Calls for AI Slowdown
5 hari lalu
An increasingly heated debate about the pace of AI development has dominated the industry. South Korea said it "cannot afford" to slow down.
Anthropic to Set up Singapore Office amid Regional AI Push
6 hari lalu
AI firm Anthropic is set to open an office in Singapore this upcoming October, marking its first hub in Southeast Asia.
Why the Era of Cheap Government Debt is Over
6 hari lalu
US debt has topped $40 trillion. How risky is the country's growing debt burden. And will investors keep funding Washington's deficits?
