Trump plan to test AI models has a problem—US security teams were gutted by DOGE (arstechnica.com)

Donald Trump recently signed an executive order aimed at enhancing the government’s approach to the voluntary safety testing of advanced AI models. However, industry critics argue that the order falls short of meaningful oversight, offering only superficial assurances about monitoring AI risks without imposing any significant regulatory changes. This follows a previously planned event involving tech CEOs that was abruptly canceled due to concerns that the original draft of the order might hinder AI innovation, reflecting ongoing tensions within the administration regarding the balance between regulation and technological advancement. The new order introduces a voluntary framework for AI companies to engage in safety reviews with the government but does not mandate any requirements or standards. Critics believe this approach may lead to insufficient oversight of potentially dangerous AI deployments. By emphasizing a commitment to “America First” cybersecurity efforts while prioritizing rapid deployment of AI technologies, the executive order raises questions about its effectiveness in ensuring safety amid evolving AI capabilities and risks. As the AI/ML community grapples with the implications of this order, it highlights the ongoing debate over the need for balanced regulation that fosters innovation without sacrificing safety.