Trump Signs Executive Order on AI Security

Washington DC. US President Donald Trump has signed a new executive order on artificial intelligence (AI). The order urges the US government to have early access to advanced AI models before they are made publicly available. According to the government, the purpose of this step is to assess cybersecurity risks and protect the country's critical infrastructure from potential AI-based cyberattacks.

The executive order requests AI companies to voluntarily share advanced cyber capabilities with the government 30 days before making new models available to other partners or customers. This order comes especially after the public release of a model called Mythos, developed by domestic AI company Anthropic. US government and financial sector officials became concerned after it was claimed that the model could detect and exploit cybersecurity vulnerabilities at an unprecedented speed.

A preliminary draft of the order prepared in May had proposed a 90-day government review. However, many companies suggested that a shorter review period would be appropriate due to the rapid development of AI technology. Subsequently, the review period was limited to 30 days in the final order.

A White House spokesperson described the order as an example of the Trump administration's practical approach to balancing innovation and security. According to him, this will further strengthen America's global leadership in the field of AI and cybersecurity.

The order also includes a provision for national security agencies to establish a cybersecurity clearinghouse to strengthen cybersecurity. However, the order clearly states that access to AI models before public release will be voluntary. It has been stated that this cannot be used as a basis for mandatory licensing, prior approval, or government permission systems for AI companies.

Major AI companies held direct discussions with the White House during the process of drafting the order. Among them are OpenAI, Anthropic, and Microsoft. According to experts, advanced AI models can significantly increase the capability of cyberattacks. Therefore, government evaluation of such systems before they are made public can help in identifying potential threats and preparing defenses.