{"slug": "amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions", "title": "Amazon CEO reportedly raised concerns before Anthropic restrictions", "summary": "Amazon CEO Andy Jassy reportedly raised security concerns about Anthropic's AI models with White House officials, leading the U.S. government to issue an export control directive that forced Anthropic to disable its Fable 5 and Mythos 5 models globally. The directive, triggered by a report showing Amazon researchers jailbroken portions of Mythos, ordered suspension of access for foreign nationals, including Anthropic's own employees. Anthropic disputed the severity of the jailbreak, calling it a minor vulnerability that other public models also exhibit.", "body_md": "# Amazon CEO reportedly raised concerns before Anthropic restrictions\n\nAmazon reportedly triggered the U.S. government's decision to disable Anthropic's Fable 5 and Mythos 5 models on Friday. According to Axios, Amazon called senior administration officials Thursday night to share a report showing its researchers had jailbroken portions of Anthropic's Mythos model using cybersecurity queries. At least five other companies also contacted officials, per Axios. Anthropic received a Commerce Department export control directive at 5:21 p.m. ET on June 12 ordering suspension of all foreign-national access, including Anthropic's own foreign-born employees. Unable to filter users by nationality in real time, Anthropic disabled both models globally. Anthropic disputed the severity, saying the technique consists of asking the model to review a codebase for flaws - a task achievable with other public models - and that no universal jailbreak has been found. The Wall Street Journal reported that Amazon CEO Andy Jassy raised the security concerns with White House officials ahead of the directive, while White House AI adviser David Sacks confirmed the administration asked Anthropic CEO Dario Amodei to fix the issue or take the model down.\n\n### Background\n\nFable 5, Anthropic's general-use version of its more advanced Mythos 5 model, launched on June 9, 2026. Anthropic said it had notified U.S. government officials multiple times before the release, and the government did not object, per Axios. The models had been publicly available for roughly three days before being pulled.\n\n### What triggered the directive\n\nAccording to Axios, Amazon called senior administration officials on the evening of June 11 to share a report showing how its researchers had jailbroken portions of Anthropic's Mythos model, accessing information about software vulnerabilities. At least five other companies also contacted administration officials, per Axios. The Wall Street Journal separately reported that Amazon CEO Andy Jassy raised security concerns about Anthropic's models directly with White House officials ahead of the government action; Semafor reports The Information first disclosed Jassy's direct contact.\n\n### China-linked access concern\n\nSemafor reported exclusively on June 13 that the White House imposed the export controls partly over suspicions that a China-linked group had accessed Mythos. An Anthropic spokesperson said the White House did not raise Chinese access to Mythos in its conversations around the Fable jailbreak and the export controls. The company prohibits access to its products from within China. It is unclear which organization accessed the model or how it gained access, per Semafor.\n\n### The Commerce Department order\n\nAnthropic received an export control directive at 5:21 p.m. ET on June 12, ordering suspension of Fable 5 and Mythos 5 access for any foreign national, whether inside or outside the United States - including Anthropic's own foreign-national employees. Unable to filter users by nationality in real time, Anthropic disabled both models for all customers globally, per the company's public statement.\n\n### Anthropic's position\n\nAnthropic stated the demonstrated jailbreak \"essentially consists of asking the model to read a specific codebase and fix any software flaws\" - a capability, the company said, available in other publicly deployed models including OpenAI's GPT-5.5. Anthropic described the findings as \"minor vulnerabilities\" offering \"no Mythos-specific uplift\" and said no universal jailbreak has been found. Luta Security CEO Katie Moussouris, who Anthropic shared the Amazon report with, told Axios the researchers \"were able to find security vulnerabilities by asking questions normal defenders would ask AI, which is exactly what the model was intended to do.\"\n\n### Government and industry reactions\n\nWhite House AI adviser David Sacks alleged on social media that when the administration notified Anthropic of the jailbreak, CEO Dario Amodei said it was not a serious risk and refused to fix it; Sacks said the administration then reluctantly issued the export controls. Former Trump administration AI adviser Dean Ball called the move \"baffling,\" citing the inconsistency with the administration's recent loosening of chip export controls to China, per TechPolicy Press.\n\n### Broader implications\n\nAn unnamed person familiar with the matter told Axios the episode amounts to \"a de-facto licensing regime,\" and that \"companies will not screw with the White House.\" An administration official told Axios that other models are not currently viewed as national security threats because they do not surpass the capability bar set by Mythos - suggesting any future frontier model at that level would face similar government scrutiny before deployment. Anthropic warned that if the same jailbreak standard were applied across the industry, it \"would essentially halt all new model deployments for all frontier model providers.\"\n\n## Scoring Rationale\n\nAmazon's reported role as triggering party raises significant conflict-of-interest questions given its major investment in Anthropic, and Semafor's exclusive reporting on China-linked access suspicions adds a geopolitical national security dimension. Anthropic's warning that the same standard would halt all frontier deployments makes this a defining development for AI governance, warranting a strong score for its policy and deployment implications.\n\nPractice with real Retail & eCommerce data\n\n90 SQL & Python problems · 15 industry datasets\n\n250 free problems · No credit card\n\n[See all Retail & eCommerce problems](/problems/datasets/retail)", "url": "https://wpnews.pro/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions", "canonical_source": "https://letsdatascience.com/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restr-d6c5990a", "published_at": "2026-06-15 06:12:47.658168+00:00", "updated_at": "2026-06-15 06:12:50.605845+00:00", "lang": "en", "topics": ["ai-safety", "ai-policy", "large-language-models", "ai-research"], "entities": ["Amazon", "Anthropic", "Andy Jassy", "David Sacks", "Dario Amodei", "Commerce Department", "White House", "Axios"], "alternates": {"html": "https://wpnews.pro/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions", "markdown": "https://wpnews.pro/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions.md", "text": "https://wpnews.pro/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions.txt", "jsonld": "https://wpnews.pro/news/amazon-ceo-reportedly-raised-concerns-before-anthropic-restrictions.jsonld"}}