Europe forces Android and Google Search open just as agentic AI shows its operating risk

Facts July 19, 2026 · Reading July 19, 2026

The European Commission issued binding measures to Google under the Digital Markets Act that require Android to give rival AI assistants equal access to key operating-system functions such as voice activation and in-app actions, and require Google to share anonymised search data with third-party search engines and AI chatbots through an API on fair, reasonable and non-discriminatory terms with privacy safeguards. Hugging Face said it detected and contained an intrusion into part of its production infrastructure that was executed end-to-end by an autonomous AI agent exploiting dataset-processing code-execution paths, and said it closed the vulnerable paths, rebuilt nodes, rotated credentials and brought in external forensics and law enforcement. These two moves together are likely to narrow two advantages that have helped the largest platforms most: control of the interface and control of the feedback data.

The Commission also accepted X’s action plan under the Digital Services Act, including commitments on API access and free researcher access to public data subject to external audit. That is likely to matter beyond X, because it suggests European regulators are not treating platform data as an internal asset by default when public oversight or downstream competition depends on it. For Microsoft, this appears to sharpen the point that distribution and trust are being reordered at the same time: the operating-system layer is becoming harder to keep exclusive, while the costs of letting autonomous agents act inside live systems are becoming harder to wave away. The effect is likely to be broader than one company, because generative AI is moving from a contest over model quality alone to a contest over who gets access, who gets data and who can operate safely enough to be allowed into everyday workflows.