Trump signs the Presidential Ordinance before publishing the AI model

On Tuesday, President Donald Trump signed an Implementing Regulation to create a “voluntary framework” that enables AI companies to share their border models with the federal government before they are released, “to promote safe innovation and strengthen the cybersecurity of critical infrastructure.” Trump signed an Implementing Regulation to review AI models before their publication

However, it is up to the AI companies to decide whether they share information with the government. However, it is up to the AI companies to decide whether they share information with the government. The order states that the US AI industry has had some success, “because we refuse to suffocate this innovation through excessive regulation,” but they also acknowledge that new AI capabilities are associated with security risks. Accordingly, it points to several federal authorities to develop a framework to “assess the advanced cyber skills of AI models” before making them accessible to the public. It is at the discretion of companies to share their models with the government's pre-publication, but could use certain confidentiality measures if they so wish. In addition, the Federal Government must prepare cyber defence measures against AI, in particular for critical infrastructures. The order came after Trump had postponed the signing of a previously planned implementing regulation at the last minute, which he feared could “impose” the competition with China. While the previous version would have allowed AI companies to freely share their models 14 to 90 days prior to publication, the current version according to the New York Times requires companies to share their models up to 30 days prior to public publication. Google, Microsoft and xAI have agreed last month to allow a pre-examination by the Center for AI Standards and Innovation (CAISI) of the Ministry of Commerce. OpenAI and Anthropic had already agreed with President Joe Biden in 2024 to share their models with CAISI as part of Biden's advance for AI security planks. But until recently, the Trump administration under the former AI group of the White House, David Sack, played down security concerns and pursued a reluctant approach. The order signed on Tuesday expressly states that this should not be regarded as a form of compulsory licensing or pre-authorisation. Nevertheless, it reflects a certain readiness of the Trump administration to take over the supervision of AI companies. “The White House is officially full of myths”

One factor for this shift could be the limited introduction of the powerful myth model by Anthropic in April, which, according to the company’s information, had reported “turning off vulnerabilities with high severity, including some in all major operating systems and web browsers”. Mythos also seemed to create a way to reduce the tensions between Anthropic and the government after there had been a dispute with the Pentagon about the use of AI for autonomous deadly weapons and mass surveillance. To date, the newly signed regulation has even been praised by groups that have opposed restrictions on state AI laws. “The White House is officially full of myths,” said Brad Carson, President of Americans for Responsible Innovation, in a statement, adding that the order shows that the Trump administration takes AI security. Brendan Steinhauser, CEO of the Alliance for Secure AI, said his group was “joyed to see that the Trump administration takes the risks of these models seriously”, and both Steinhauser and Carson demanded the Congress to codify binding protective measures. Most popular

- Gemini Spark is the most impressive and frightening AI experience I have done so far

– The new AI agent from Gemini is about as good as the demo of Google

– These are the first Nvidia RTX Spark laptops

– This could be the M1 moment of Windows be – but it is expected to cost a lot

– This is the Microsoft Surface Laptop Ultra with Nvidia RTX Spark