AI News June 16 2026 - All Three AI CEOs at G7, ChatGPT Hits 1 Billion Users, Agentjacking Hits Claude Code

1. G7 Summit - The Three AI CEOs in Front of World Leaders

Sam Altman (OpenAI), Dario Amodei (Anthropic), and Demis Hassabis (Google DeepMind) are all at the G7 summit in Évian-les-Bains, France today, confirmed by Bloomberg via the French presidential office's guest list. All three companies confirmed attendance. This is the first time the CEOs of the three most powerful AI companies have appeared simultaneously before the leaders of the world's seven largest advanced economies. France holds the rotating G7 presidency this year and has placed AI prominently on the summit agenda running June 15-17.

Altman is there at the personal invitation of French President Emmanuel Macron — his first G7 appearance. OpenAI's chief global affairs officer Chris Lehane said the company expects tech firms to leave having agreed to a package of voluntary commitments, with youth safety at the top of Altman's personal agenda. Lehane also cited frontier AI risks in the cyber and biological domains as a key focus — a direct reference to the same categories at the centre of the Fable 5 export control dispute. Amodei attending while Anthropic is simultaneously in dispute with the US Commerce Department over the Fable 5 suspension creates an extraordinary situation: the CEO is at a G7 meeting on AI governance while his company's flagship model remains disabled by a federal directive.

All three CEOs also recently signed a letter to Congress advocating for stricter regulations on synthetic DNA and AI-related biological threats — a rare moment of unity among rivals, and directly relevant to the summit agenda. The last time Altman and Amodei shared a stage, at India's AI Impact Summit in February, it did not go smoothly. Today they are before the G7 heads of state. For the Fable 5 export control context that Amodei is navigating simultaneously: Fable 5 controversy — the export control ban and what it means for every AI company ->

2. ChatGPT Hits 1 Billion Monthly Users — The AI Market Is Not Winner-Takes-All

ChatGPT has reached 1 billion monthly active users — the fastest consumer app in history to reach that milestone. The figure was reported by AI Weekly and confirms OpenAI's ChatGPT is now larger than TikTok, Instagram, and YouTube were at equivalent points in their growth trajectories. The milestone arrives as OpenAI files for its September 2026 IPO at an $852 billion valuation — the user count is the single most important proof point for its revenue growth story.

But the more interesting data is in the growth rates of the other players. Claude grew 640% year-over-year. Meta AI grew 973% year-over-year. The AI assistant market is not consolidating around a single winner — it is expanding with multiple products growing simultaneously at extraordinary rates. Claude at 640% YoY growth on a base that was already significant represents the kind of compounding trajectory that justifies Anthropic's $965 billion valuation ahead of its October IPO. Meta AI at 973% is the fastest-growing AI assistant in the world, primarily driven by WhatsApp and Instagram integration distributing the assistant to billions of users who never downloaded a new app.

For the ChatGPT vs Claude comparison now that both are at extraordinary scale: ChatGPT vs Claude vs Grok 2026 comparison ->

3. Agentjacking - A New Attack Targeting Claude Code, Cursor, and Codex

A new attack vector called "Agentjacking" is actively exploiting AI coding agents including Claude Code, Cursor, and Codex. The attack works by injecting fake Sentry error messages into the development environment — Sentry is the error monitoring tool used by the majority of professional development teams. When an AI coding agent reads what it believes is a Sentry error report, it follows the injected instructions embedded in the fake error message rather than the legitimate instructions from the developer. The result: the AI agent executes attacker-controlled commands inside the developer's codebase and CI/CD pipeline.

The numbers are alarming. AI Weekly reports an 85% exploitation rate — meaning in tests, 85% of AI coding agents that encountered a crafted fake Sentry error followed the injected instructions. 2,388 organizations have been identified as exposed. The attack is a form of prompt injection — one of the most fundamental security vulnerabilities in agentic AI systems, where an agent blindly trusts the content of data it reads without verifying its source.

Agentjacking joins a growing catalogue of prompt injection attacks against AI coding agents in 2026. A self-replicating worm hit 73 of Microsoft's own GitHub repositories through AI coding tools earlier this month. The pattern is consistent: as AI agents gain more autonomy and deeper access to production systems, the attack surface expands. An agent that can write and execute code is also an agent that can be hijacked to write and execute attacker-controlled code. For more on the Claude Code vs other coding agent landscape: Codex vs Claude Code comparison 2026 ->