OpenAI has rolled out its most powerful autonomous developer capabilities to European users, but the expansion comes with strict privacy guardrails and a looming model retirement. On June 16, the company extended Computer Use, the Codex Chrome extension, Memories, and Chronicle to users across the European Economic Area, the United Kingdom, and Switzerland. These features had been unavailable in those regions since their original release months earlier. Simultaneously, OpenAI confirmed that GPT-4.5 will exit the ChatGPT interface on June 27, though the model remains available through the company's API.

What Is Computer Use and How Does It Work Differently on Mac and Windows?

Computer Use is the headline feature in this expansion. It allows Codex to see, click, type, and navigate inside desktop applications on both macOS and Windows without requiring any special integration or plugin. The technology operates through fundamentally different architectures on each platform, with real consequences for how developers can use it.

On macOS, Codex leverages Apple's private APIs to send input events like clicks and keystrokes to background windows while the developer continues working in the foreground. A local HTTP server on the machine accepts these commands and dispatches them to target applications. Multiple Codex agents can run simultaneously in the background without interfering with what the user sees on screen. Screenshot data is processed locally where possible, while metadata including labeled interface elements, text extracted through optical character recognition (OCR), and action logs are sent to OpenAI's cloud for reasoning, encrypted during transit. Enterprise deployments can route this reasoning through Azure OpenAI Service on-premises instead.

Windows presents a different trade-off. Codex runs in the foreground and takes over the active desktop session for the duration of a task, meaning a developer cannot work on the same machine while Codex is using it. OpenAI sandboxes both implementations using technologies like bubblewrap on Linux and Windows Sandbox on Windows to limit what the agent can write outside its working directory. This sandboxing reflects a deliberate engineering acknowledgment that granting an AI agent unrestricted desktop access introduces real, not hypothetical, risk.

Why Are Memories Turned Off by Default in Europe?

The Memories feature allows Codex to retain user preferences, recurring workflows, technology-stack choices, and repository conventions across sessions so that future interactions require less setup. However, in the European Economic Area, this feature arrives with a critical configuration: it is off by default, and users must actively enable it. This design choice is not arbitrary; it reflects compliance with the General Data Protection Regulation (GDPR).

Under GDPR, a system that builds persistent behavioral profiles of users constitutes profiling activity under Article 4, which triggers obligations around consent, data minimization, and the right to erasure. OpenAI Ireland Limited, the registered EEA data controller based in Dublin, made the off-by-default decision to align with GDPR's data minimization principle, which requires that personal data be limited to what is necessary for the purposes for which it is processed. However, whether passive opt-in, where users activate Memories by switching it on, satisfies GDPR's consent standard remains contested. European data protection authorities have previously questioned whether feature defaults in consumer AI products meet the bar for "freely given, specific, informed and unambiguous" consent.

Italy's Garante fined OpenAI 15 million euros in December 2024 for GDPR violations related to ChatGPT's data processing practices, establishing that European regulators are prepared to use enforcement tools against AI products. For users who do enable Memories, the data that Codex retains may be subject to deletion requests under GDPR Article 17. OpenAI's Europe privacy policy states that personal data marked for deletion is removed from its systems within 30 days.

How to Prepare for the EU AI Act's Enforcement Deadline

• Understand the August 2 Deadline: The EU AI Act's enforcement powers become fully operational on August 2, 2026, enabling fines of up to 15 million euros or 3% of global annual turnover for general-purpose AI (GPAI) non-compliance, alongside transparency obligations requiring providers to disclose when users interact with AI.
• Review Documentation Requirements: OpenAI must maintain technical documentation, publish training data summaries, and adopt copyright compliance policies under the GPAI model obligations that took effect August 2, 2025.
• Evaluate Your Use Cases: Computer Use and Memories represent some of the most consequential autonomous behavior offered in a commercial AI developer platform, and both fall within the EU AI Act's GPAI category, adding to the transparency and oversight documentation OpenAI must maintain.

The Codex EEA expansion arrived 44 days before this concrete enforcement threshold. The timeline makes the European rollout a deliberate positioning move: European developers now have weeks to build with these capabilities before the regulatory environment surrounding them becomes stricter. A May 2026 provisional political agreement between EU institutions extended high-risk AI system compliance deadlines, deferring some obligations to December 2027, but left the August 2, 2026 enforcement date for GPAI obligations and Article 50 transparency requirements unchanged.

What Happens to GPT-4.5 After June 27?

OpenAI confirmed in its ChatGPT release notes that GPT-4.5 will be retired from the ChatGPT interface on June 27, 2026. This retirement applies exclusively to the ChatGPT interface and does not affect API access to any OpenAI models, including GPT-4.5 and o3. The distinction is crucial for developers and enterprises: API access to all OpenAI models remains available through the API with no current retirement date. The retirements announced apply only to the ChatGPT interface, not to any programmatic integration.

For ChatGPT users accustomed to GPT-4.5, the model will no longer be available as a selectable option in the web interface after June 27. However, developers and organizations using OpenAI's API can continue to call GPT-4.5 and other models without interruption. This separation between the consumer ChatGPT product and the developer API reflects OpenAI's strategy of managing different product lifecycles independently.

The timing of these announcements underscores OpenAI's dual focus: expanding autonomous capabilities in regulated markets while managing model availability across different product tiers. European developers now have access to some of the company's most powerful tools, but within a framework designed to satisfy emerging regulatory requirements. Meanwhile, the broader developer ecosystem retains stable API access to GPT-4.5 and newer models, ensuring that production systems built on these models will not face forced migrations.