Why Microsoft Blocked Employees From Using Anthropic’s Claude Fable 5 Despite Offering It to Customers

Microsoft has reportedly restricted its employees from using Anthropic's newly launched Claude Fable 5 AI model, citing concerns over the company's updated data retention policies. The move comes despite Microsoft simultaneously making the same model available to enterprise customers through GitHub Copilot and Microsoft Foundry, highlighting the complex compliance challenges surrounding advanced AI adoption.

Why Has Microsoft Blocked Claude Fable 5?

According to reports, Microsoft has disabled access to Claude Fable 5 within the internal GitHub Copilot model picker used by its employees. The restriction follows Anthropic's decision to introduce a new data retention framework for its powerful Mythos-class AI models. Microsoft Brings Claude Models to Foundry, Expands Azure’s Access to Frontier AI Tool.

Anthropic launched Claude Fable 5 on June 9, describing it as the most capable public version of its Mythos AI series. The company said the model includes safeguards that prevent its use in sensitive and high-risk areas, including certain cybersecurity applications.

What Changed in Anthropic's Data Retention Policy?

Under Anthropic's revised policy, prompts submitted to Claude Fable 5 and the model's generated responses are retained for 30 days for trust and safety monitoring purposes. Additionally, if Anthropic's automated safety systems identify content that may violate platform policies, the associated inputs and outputs can be stored for up to two years. Microsoft HR Chief Amy Coleman Promises Greater Transparency Amid Employee Concerns; Read Full Internal Memo.

The policy applies across both Anthropic's own platforms and third-party services that offer Mythos-class models.

For Microsoft, which routinely handles sensitive customer information, proprietary software code, and confidential enterprise data, the retention period has reportedly triggered legal and compliance concerns. The company's legal teams are evaluating whether the new requirements align with Microsoft's internal privacy, security, and data governance standards.

Other Claude Models Remain Available

The restriction currently applies only to Claude Fable 5. Other Claude models continue to be accessible to Microsoft employees because they operate under Anthropic's Zero Data Retention framework, which offers stronger privacy guarantees for enterprise users.

Anthropic's Response

Anthropic has defended the policy, stating that retained data is used exclusively for trust and safety purposes and is deleted after 30 days in nearly all circumstances. The company also emphasized that the retained prompts and outputs are not used to train future Claude models and are not repurposed for commercial objectives.

A Growing Enterprise AI Dilemma

The situation underscores a broader challenge facing enterprises adopting generative AI technologies. A model can be commercially available and integrated into customer-facing products while still failing to meet the internal compliance standards of the company distributing it.

As organizations increasingly balance AI innovation with privacy, security, and regulatory obligations, the Microsoft-Claude Fable 5 standoff illustrates how data governance policies can become just as important as model performance in determining enterprise adoption.

(The above story first appeared on LatestLY on Jun 11, 2026 04:36 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).