• Trump's executive order mandates voluntary sharing of AI models for cybersecurity review.
• The order allows the government to vet AI systems for up to 30 days prior to public release.
• Companies must cooperate to enhance national security without mandatory regulations.
• Concerns over AI models like Anthropic's Mythos prompted the change in approach.
• This move aims to preserve US leadership in AI technology amid competition with China.

On June 2, President Trump signed an executive order focused on enhancing cybersecurity through artificial intelligence (AI) by establishing a voluntary framework for AI companies to share advanced models before they're released to the public. The order permits federal agencies to evaluate these models for up to 30 days, aiming at balancing national security interests with innovation, as expressed by Trump: "We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead," according to CBS News, Los Angeles Times, and South China Morning Post.

The order represents a shift from the administration's previous, more lenient stance towards AI regulation, now aiming to mitigate potential national security risks posed by powerful AI systems such as Anthropic's Mythos. Trump's administration plans to secure agreements with AI companies, encouraging them to voluntarily submit their models for cybersecurity assessments, which could address vulnerabilities that pose risks to critical infrastructure, according to NPR, Reuters, and CBS News.

The new framework stops short of enforcing mandatory regulations on AI companies, emphasizing that no mandatory governmental licensing or preclearance will be imposed on new AI models, thus maintaining a balance between security needs and innovation freedom. The goal is to foster collaboration with AI leaders like OpenAI and Google without stifling technological advancements in the U.S., according to Los Angeles Times, Reuters, and NPR.

The administration's decision follows heightened anxiety regarding the release of AI technologies that could exploit software vulnerabilities, leading to this new operational paradigm. With concerns amplified by the capabilities of models like Mythos, which can identify critical software flaws, the objective is to navigate the competitive landscape of AI without compromising national safety, as outlined in statements from sources including South China Morning Post and Los Angeles Times.