Apple picks Gemini to rebuild Siri, ChatGPT crosses 1.1 billion, and Congress drafts its first AI law
Apple's WWDC AI pivot to Google Gemini, OpenAI's 1.1B-user milestone, a bipartisan US AI bill, and Europe's August deadline.
By BINA Editorial
Apple has chosen Google Gemini to power a rebuilt Siri, confirming a strategic bet announced at WWDC earlier this month. ChatGPT just crossed 1.1 billion monthly users — the fastest any app has ever reached that figure — yet OpenAI's share of the AI assistant market is shrinking. On Capitol Hill, a bipartisan pair of representatives have published the country's first comprehensive federal AI governance bill, a 269-page discussion draft that would preempt state-level AI legislation for three years. And with six weeks remaining until the EU AI Act's high-risk enforcement deadline, compliance teams at companies operating in Europe are discovering that preparation timelines were badly underestimated.
Apple rebuilds Siri on Google Gemini
The deal was confirmed at WWDC on June 8 and is now the clearest illustration of the competitive reset in AI assistants. Apple renamed its voice assistant "Siri AI" and rebuilt it on generative AI foundations, with Google Gemini powering the experience through a landmark multi-billion-dollar partnership — one that gives Google its most prominent consumer AI deployment outside its own ecosystem.
The move resolves a long-standing problem for Apple: Siri's limitations compared to ChatGPT and Claude had become a recurring complaint among developers and users. Rather than build a frontier model from scratch, Apple chose a licensing approach that mirrors how it handled early Maps data and how it still handles default search through Google. The company is betting that hardware, privacy architecture, and on-device integration — not model ownership — are the defensible advantages worth holding.
For Google, the deal is a significant win. Gemini is now embedded in the world's most valuable consumer hardware franchise and will reach hundreds of millions of iPhone, iPad, and Mac users who might never have opened the Gemini app independently. The financial terms have not been disclosed, but analysts familiar with the deal structure describe it as comparable in scale to Apple's existing default search agreement with Google.
The partnership also aligns Apple and Google against a common set of competitors — a shift in incentive structure that has not gone unnoticed in Redmond or San Francisco.
ChatGPT hits 1.1 billion users as rivals accelerate
OpenAI confirmed in early June, citing Sensor Tower data reported by Reuters, that ChatGPT had become the fastest consumer application in history to reach 1 billion monthly active users — outpacing Facebook, YouTube, TikTok, and Instagram on time-to-scale. The app launched on iOS and Android in early 2023 and reached the milestone in roughly three years. As of mid-June, the number has grown to 1.1 billion.
The headline obscures a more complicated competitive picture. ChatGPT's share of the AI assistant market has fallen from over 50 percent in January 2026 to 46.4 percent in May. Google Gemini and Anthropic Claude are growing at faster rates, and the total addressable market is expanding quickly enough that OpenAI can add tens of millions of users in absolute terms while losing share in relative terms.
The timing coincides with a constrained moment for OpenAI. The company has delayed its IPO timeline amid ongoing governance discussions around its for-profit conversion, which means near-term narrative is driven by product velocity rather than capital markets. GPT-5.6 — the next iteration in OpenAI's six-week flagship cadence, targeting a 1.5-million-token context window and improved agentic execution — is expected to ship this week alongside a redesigned ChatGPT interface.
Congress tables its first comprehensive federal AI bill
On June 4, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) published a 269-page bipartisan discussion draft of the Great American Artificial Intelligence Act of 2026. It is the first congressional attempt at a comprehensive federal AI governance framework, and its two most consequential structural choices concern scope and preemption.
The bill targets "large frontier developers" — companies with over $500 million in annual revenue that have trained a frontier AI model. Under Section 111, these companies would be required to publish a frontier AI framework documenting their approach to catastrophic risk mitigation. Section 112 would create a federal licensing system in which the Director of the Center for AI Standards and Innovation (CAISI) oversees "independent verification organizations" that frontier developers must retain for audits and assessments. CAISI would be formally established inside the Commerce Department and funded with $300 million over three years.
On preemption, the draft takes a selective approach: states retain authority to regulate how AI systems are used within their borders, but would lose authority over how those systems are built. A three-year preemption period on AI development regulation would effectively freeze the patchwork of state-level AI developer rules that the major labs have been navigating since Colorado passed the first such law in 2024.
The draft also includes federal whistleblower protections for employees and contractors who report AI violations — a provision notable for its scope and for being embedded in a bill otherwise focused on large enterprises.
The discussion draft is not yet legislation. Stakeholder feedback is being solicited, and formal introduction would follow in a later session. The political window for passage before the midterms is narrow, but the draft has already established the terms of the federal AI governance debate in a way that will be difficult to ignore.
Six weeks to the EU AI Act's high-risk enforcement deadline
August 2, 2026 is the enforcement date for the EU AI Act's high-risk AI system obligations under Articles 9–17, covering provider requirements including quality management systems, risk management frameworks, technical documentation, conformity assessments, and EU database registration. Article 26 deployer requirements take effect on the same date, along with Article 50 transparency obligations requiring AI chatbots to disclose their artificial nature and emotion recognition systems to notify users.
Compliance readiness is poor. Research published this month by Gartner and the Cloud Security Alliance found that a significant share of enterprises had not completed AI system inventories — the prerequisite for every downstream compliance task. The recurring explanation is that organizations treated August as a future planning item into early 2026 and then discovered in Q2 that the technical work required could not be compressed into weeks.
US companies with EU market exposure are directly in scope regardless of where their systems are built or where their headquarters is located. Penalties for non-compliance reach €15 million or 3 percent of global annual turnover, whichever is higher. Six weeks is not enough time to build a compliance program from scratch; it is, at best, enough time to complete an inventory and scope which systems require immediate remediation versus a documented exception pathway. Compliance counsel at several large American technology companies have confirmed that preparation intensity has escalated to crisis-response levels.
The EU's August deadline and Congress's still-theoretical federal framework represent two very different speeds of AI governance. One is six weeks away with binding penalties; the other is a discussion draft without a legislative calendar. Companies operating across both jurisdictions are managing that asymmetry in real time.