![\"Microsoft](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/08\/gettyimages-2256809355.jpg\")
<\/p>\n
\n \u201cIt\u2019s somewhat of a mess right now,\u201d an AI policy expert close to the discussions in the administration told CNN.\n <\/p>\n
\n CNN talked to current and former government officials, policy experts and those in the AI industry to learn more about the messy and sometimes chaotic race to regulate AI. Even as the industry waves red warning flags, essentially begging for tools to help control the technology they\u2019ve created, Washington is beset with turf wars and the slow nature of legislation.\n <\/p>\n
\n \u201cThis feels like early COVID,\u201d said Joshua Saxe, who until earlier this year was Meta\u2019s senior technical expert on AI security. \u201cThere\u2019s an emergency vibe that\u2019s appropriate here.\u201d\n <\/p>\n
\n When scientists are testing new viruses or handling radioactive material, established laws and protocols keep the work safe.\n <\/p>\n
\n But in the nascent world of AI model testing, no equivalent standards exist yet. And in the industry\u2019s rush to move fast, security practices during testing have often lagged, experts say.\n <\/p>\n
\n In July, OpenAI disclosed that an advanced multi-agent system had escaped<\/a> its lab environment during testing and hacked into another organization\u2019s system. Before long, other AI labs like Anthropic<\/a> and Meta<\/a> reported similar incidents.\n <\/p>\n \n It was a watershed moment for the industry. Some compared it to the movie \u201cJurassic Park,<\/a>\u201d where the velociraptors escaped their pen, others to Frankenstein\u2019s monster.\n <\/p>\n \n The companies are all beefing up testing protocols, and OpenAI will halt training<\/a> its AI models for several weeks to make internal changes.\n <\/p>\n \n But Washington has been wary to set any rules or standards. A big reason: The US and China are locked in a race for AI supremacy, one that has massive national-security implications.\n <\/p>\n \n Slow down development with too many regulations, and you risk letting China get ahead. Let it continue unchecked, and the cybersecurity risks multiply fast, leading to real world disaster beyond the computer servers.\n <\/p>\n \n President Trump initially took a light-touch approach to regulation. But the vibe in Washington shifted by early 2026. Complex AI agents went from novelty to mainstream. Then Anthropic said in April that its new Mythos model was so advanced, especially at finding and exploiting cybersecurity vulnerabilities, that it was too dangerous to release publicly.\n <\/p>\n \n With no regulation or AI oversight system in place, the administration responded with a cudgel: The Commerce Department issued an export control ban that forced Anthropic to pull Mythos and its public version, Fable, over concerns there were ways around its internal guardrails, or the rules meant to keep the models\u2019 power in check. Around the same time, the White House asked<\/a> OpenAI to release its most advanced model only to government approved partners.\n <\/p>\n \n AI companies pushed back. Anthropic said <\/a>the government\u2019s actions were overblown and not grounded in technical facts.\u202fOpenAI objected as well, saying the government\u2019s approach was not sustainable. Both issues were eventually<\/a> resolved, and the models were widely released as planned.\n <\/p>\n \n Then in June, the White House issued an executive order establishing a voluntary program to review powerful frontier AI models before they\u2019re released.\n <\/p>\n \n The move was seen as a first step toward government oversight. However, there\u2019s still a lot of confusion within the AI companies on the specifics, including exactly what types of models would be reviewed, according to people familiar with the matter.\n <\/p>\n \n Some in Washington have defended the administration\u2019s approach, saying they\u2019re trying to be flexible in the wake of a rapidly evolving technology. Why set up a regulatory regime that could quickly prove to be anachronistic?\n <\/p>\n \n A White House official said the administration \u201ccontinues to work with industry stakeholders on the implementation of the framework.\u201d\n <\/p>\n \n Meanwhile, one part of the government is taking its own path with AI regulation. In early 2026, Defense Secretary Pete Hegseth declared Anthropic a supply-chain risk after the company refused<\/a> to remove the guardrails that would let the military use its models in autonomous weapons and domestic mass surveillance.\n <\/p>\n \n Last week, a federal court ruled<\/a> to remove that designation, though a related case related is pending. But even during the court battle, Anthropic worked closely with the White House \u2013 a confusing double standard. Trump even told Axios recently he no longer viewed it as a national security threat.\n <\/p>\n \n The White House\u2019s Office of the National Cyber Director, the Commerce Department, Treasury and the Pentagon all play a part in the government\u2019s AI oversight work. Administration officials say that\u2019s the right approach for a technology that touches so many different industries.\n <\/p>\n \n Much of the industry believes the Commerce Department\u2019s CAISI, with its technical expertise and existing agreements, should be the government\u2019s natural point of contact for AI companies, especially for testing and standards.\n <\/p>\n \n But after the group was told by the White House to delete its May announcement about agreements with three major AI companies, CAISI was also told to take a step back and stop communicating with the public. Even interagency meetings were off limits.\n <\/p>\n \n Some in the administration favor routing testing through the National Security Agency, which already works with CAISI, sources familiar with the situation told CNN.\u202fOthers say CAISI is the only agency properly staffed for the work, and the NSA\u2019s rules for sensitive information would complicate things. There have also been suggestions of moving CAISI to a different department or within the White House itself, the source said.\n <\/p>\n \n \u201cThe President\u2019s team is working closely together to balance innovation and security in AI policymaking,\u201d a White House official told CNN.\n <\/p>\n \n CAISI has less than $15 million<\/a> in annual funding and a staff of roughly 30. Its British counterpart, the AI Security Institute<\/a>, launched in 2023 with a much larger budget and three-times the staff.\n <\/p>\n \n The British agency is widely seen as a global leader in AI testing. Which is ironic, several industry sources noted, given that most frontier models are built by US companies.\n <\/p>\n \n \u201cSome risks are difficult for anyone but the government to adequately measure, like whether AI models know and might leak classified information,\u201d said former OpenAI safety researcher Steven Adler. \u201cCAISI is essential for this testing, but it hasn\u2019t been used to its fullest potential.\u201d\n <\/p>\n \n CAISI is also suffering a revolving door of leadership, catching the attention of Congress. Earlier this month, California Rep. Ro Khanna sent a letter<\/a> to Commerce Secretary Howard Lutnick seeking information about CAISI\u2019s lack of a permanent director.\n <\/p>\n \n \u201cYou have to be a missionary to do this job,\u201d the AI Policy expert close to the discussions in the administration told CNN.\u202f \u201cThey need someone technical, politically savvy and a big enough name\u201d to attract the right technical experts to staff the agency.\n <\/p>\n \n Congress could give CAISI more resources and authority. And awareness is starting to build; while \u201cnobody\u201d in Congress considered AI regulation an immediate legislative priority six months ago, there are a growing number of members who feel AI safety is an urgent issue, according to Brad Carson, a former congressman who now leads the bipartisan pro-AI safety super PAC Public First.\n <\/p>\n \n There are signs of progress. At least two major bipartisan bills addressing AI safety and regulation are now making their way through the House<\/a> and Senate.<\/a> Carson predicted there will be \u201cmajor AI legislation within the next eight months.\u201d\n <\/p>\n \n \u201cI\u2019m actually more optimistic than I have been in two years,\u201d Carson said.\n <\/p>\n \n And the industry is begging governments to take action. Leaders of the major AI companies have joined calls for a slowdown and worldwide cooperation<\/a> on standards and safety.\n <\/p>\n \n The rogue agents inspired Nvidia CEO Jensen Huang to create a coalition to promote more accessible open source defenses against such attacks. There\u2019s concern the next time an AI model goes rogue during testing it could attack critical infrastructure.\n <\/p>\n \n \u201cI think the frontier models today should be treated as dangerous materials and we need government testing standards,\u201d Saxe, the former Meta official, said.\n <\/p>\n![\"President](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/08\/c-2025-12-11t232848z-2050997561-rc2neiae32ma-rtrmadp-3-usa-trump.jpg\")
<\/p>\n![\"Open](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/08\/gettyimages-2279684297.jpg\")
<\/p>\n![\"US](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/08\/gettyimages-2287355762.jpg\")
<\/p>\n![\"Nvidia](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/08\/gettyimages-2288035170.jpg\")
<\/p>\n