cybersecurity risks

OpenAI is tightening controls on its new model due to cybersecurity risks, amidst an intensifying AI security debate. The AI lab warned that a new model could potentially reach "Critical" capability, allowing it to launch cyberattacks against sophisticated cyber defenses.

The White House discussed a framework for evaluating cybersecurity risks of frontier AI models with OpenAI and Anthropic but will not release details publicly. Critics argue the secrecy undermines confidence in the evaluation process, while proponents suggest it provides the US government a strategic advantage in AI development. The vetting focuses on top-tier US-developed AI products, excluding lower-cost software.

The Trump administration's AI testing framework excludes open models and cannot restrict them post-release. The framework was established after an executive order in June requiring AI companies to share frontier models with the government before release to address cybersecurity risks.

China is threatening retaliation in response to a U.S. ban on foreign humanoid robots. Beijing's commerce ministry has called on Washington to reverse an FCC decision that claims cybersecurity risks threaten U.S. supply chains.

OpenAI is set to release its most powerful AI model series after a delay caused by US government concerns over potential cybersecurity risks.

Chinese AI companies like DeepSeek and Zhipu are aggressively expanding to compete with the US in artificial general intelligence (AGI) development. While capturing a growing AI market share through open-source models, experts warn that the lack of cooperation could lead to catastrophic cybersecurity risks.

April's hacking incidents resulted in losses exceeding $600 million, with risks expanding beyond traditional code vulnerabilities. The report highlights a growing trend of cyber threats targeting broader operational and financial systems.