Leading AI firm Anthropic has got permission from the United States government to permit a select few of US cybersecurity companies to employ it’s advanced AI model namely Mythos 5. This is a partial relaxation of certain restrictions put into effect by the US Administration citing credible threats to national security. As per some reports the American government gave the order post warnings by Amazon as well as other organizations that advanced AI models were vulnerable to jailbreaking and misuse by malicious players.

Concerns over Misuse

One major issue was the potential of misuse of advanced AI models by military intelligence agencies in China, Russia and other potentially hostile entities. For your knowledge the restrictions were put earlier this month. Extensive discussions were held between Anthropic and American authorities to resolve the perceived security risks that led to the blockade.

Partial Agreement Reached So Far

US Commerce Secretary Howard Lutnick stated that the American administration was satisfied that sufficient safeguards were put in place. In a statement issued by Anthropic, the company said they were happy with the progress made so far and would work in tandem with the authorities to broaden the availability of Mythos 5 as well as restore the general use of Fable 5 soon.

Restrictions on Foreign Nationals

After withdrawing the export block made earlier in June, the American government has permitted Anthropic to select companies as well as organizations. However, permission has not yet been granted to Anthropic to release Fable a relatively less capable version of Mythos. The US authorities had earlier directed Anthropic to suspend use of both Mythos and Fable by foreign nationals even if they were Anthropic employees.

Ad Hoc Regulation of AI Sector

Earlier the US government was relatively not strict when it came to regulating the artificial intelligence sector. Their approach was to ensure American companies were ahead of other competitors and those from China in particular. Among other steps, they have also ordered OpenAI which is a major competitor to Anthropic to restrict access to its latest model.

Fear of AI System’s Capabilities

A major concern is that powerful AI systems could be employed by hackers as well as other malicious entities to detect and exploit cybersecurity vulnerabilities quickly. On Friday, Open AI released its GPT-5.6 model subject to a client-by-client vetting by the authorities.

Dispute Over Applications of Technology

In recent months, Anthropic had a rocky relationship with the Trump government. The latter was particularly miffed that the organization opposed the employment of its technology for the purpose of mass surveillance as well as autonomous weapons. As a result, the Pentagon called off contracts with the AI giant.

The American government had previously designated the AI major as a “supply chain risk” and put on the blacklist. The AI giant challenged the decision legally.

FAQs

1. Why did the US government partially relax restrictions on Anthropic?
The US government eased certain restrictions after discussions with Anthropic and assurances that adequate security safeguards had been implemented. Officials were satisfied that the risks associated with the company’s advanced AI model, Mythos 5, had been sufficiently addressed for limited use by select US cybersecurity companies.

2. What security concerns led to the original restrictions on Anthropic’s AI models?
The restrictions were imposed due to concerns that advanced AI models could be exploited through jailbreaking or misused by hackers, cybercriminals, or foreign intelligence agencies from countries such as China and Russia. There were fears that these models could accelerate cyberattacks and expose critical vulnerabilities.

3. What changes have been made under the new agreement?
The US government has allowed Anthropic to provide its advanced Mythos 5 model to a limited number of approved US organizations. However, restrictions remain on the broader release of both Mythos 5 and the less capable Fable model, particularly for foreign nationals.

4. Why is the US government increasing oversight of advanced AI models?
As AI systems become more powerful, governments are increasingly concerned about their potential misuse in cyber warfare, espionage, autonomous weapons, and large-scale cyberattacks. The US administration is therefore introducing tighter controls to balance innovation with national security.

5. How has the US government’s approach to AI regulation changed?
The US government initially adopted a relatively hands-off approach to maintain American leadership in AI. However, growing national security concerns have prompted stricter oversight, including restrictions on advanced AI models from companies like Anthropic and OpenAI to prevent misuse while allowing controlled deployment for trusted organizations.

SOURCES:-

https://www.firstpost.com/world/us-eases-restrictions-on-anthropic-permits-limited-access-to-mythos-5-for-select-cybersecurity-firms-14026664.html

https://www.news18.com/amp/world/us-approves-limited-release-of-anthropic-ai-model-after-cybersecurity-concerns-ws-l-10176264.html

https://edition.cnn.com/2026/06/26/tech/anthropic-mythos-release

https://indianexpress.com/article/technology/artificial-intelligence/anthropic-resumes-limited-mythos-5-rollout-after-us-eases-restrictions-10759685/

https://timesofindia.indiatimes.com/world/us/us-approves-limited-rollout-of-anthropics-mythos-5-ai-model/articleshow/132027393.cms

https://www.deccanchronicle.com/world/americas/us-authorizes-limited-access-to-anthropics-mythos-ai-model-1966589

https://www.nytimes.com/2026/06/26/technology/anthropic-mythos-government-restrictions.html

https://www.scmp.com/news/world/united-states-canada/article/3358579/us-eases-ban-ai-model-mythos-feared-aid-cyberattacks

https://www.yahoo.com/news/us/articles/anthropic-gets-us-governments-permission-015456919.html