Anthropic's models suspended as G7 leaders convene on AI
The US suspends foreign access to Anthropic's newest models as the G7 prepares to put AI at the centre of its Évian summit.
By BINA Editorial
A Friday afternoon directive from the US Commerce Department sent ripples through the AI world this week, as Anthropic pulled its newest models offline and world leaders prepared to make artificial intelligence the centrepiece of the G7 meeting in France.
Anthropic suspends Fable 5 and Mythos 5 under US export order
On 13 June 2026, Anthropic took its two newest AI models — Fable 5 and Mythos 5 — offline to comply with a directive from the Trump administration. US Commerce Secretary Howard Lutnick wrote to CEO Dario Amodei requiring that both models be placed under export controls, blocking access for foreign governments, companies, and individuals — including foreign nationals inside the United States. Anthropic, which had publicly launched Fable 5 just four days earlier, called the action a "misunderstanding" and said it hoped to restore access soon. The move marks the most sweeping restriction the US government has yet imposed on a commercially available frontier AI model.
G7 summit in Évian to put AI at the top of the agenda
When G7 leaders gather in Évian-les-Bains, France, from 15–17 June, artificial intelligence is expected to be the dominant topic — and for the first time, the CEOs of all three leading AI companies will be in the room. OpenAI's Sam Altman, Google DeepMind's Demis Hassabis, and Anthropic's Dario Amodei appear on the French presidential office's guest list, alongside Mistral AI's Arthur Mensch and Salesforce's Marc Benioff. French President Emmanuel Macron personally invited Altman. Leaders will discuss AI regulation, online safety for minors, and critical infrastructure, building on the Hiroshima AI Process commitments of previous years. A significant rift in regulatory philosophy — the US favouring a light-touch approach, the EU pushing for sovereignty and compliance, Canada proposing public investment — is expected to be tested around the table.
New York passes seven AI bills including chatbot safety law for children
New York's legislature concluded its 2026 session by sending seven AI-related bills to Governor Kathy Hochul, who has until 31 December to sign them. The most significant measures include the Kids Chatbot Safety Bill (S 9051), which prohibits AI companions from using features deemed unsafe for minors and creates a private right of action; the FAIR News Act (S 8451), requiring generative AI tools used in news production to disclose their use; and the AI Training Data Transparency Act (A 6578), requiring developers to publish a summary of the datasets used to train their models. New York becomes the latest US state to advance substantive AI consumer protections.
Visa embeds its payment network into ChatGPT
On 10 June, Visa announced that it has integrated its payment network directly into ChatGPT, allowing AI agents to complete purchases — not just recommend them — at any merchant that accepts Visa. Users link their cards to ChatGPT, set spending limits, and OpenAI's agents handle product discovery and transaction initiation while Visa manages authorisation and fraud monitoring. The integration is the first by a major global payments network at this scale with an AI platform, and marks a significant step toward AI agents becoming autonomous commercial actors.
German court holds Google liable for false AI-generated search summaries
A Munich court has issued a preliminary injunction against Google after two German publishers were falsely described by AI Overviews as involved in "dubious business practices" — a characterisation not supported by any of the linked articles. The Munich Regional Court ruled that Google's standard caveat — warning users to verify AI results — is insufficient to avoid liability; since only Google can adjust the models producing AI Overviews, the court treated those summaries as Google's own statements. The ruling sets a significant precedent for AI-generated content liability in the European Union.
Canada introduces Safe Social Media Act to protect children from AI chatbots
On 10 June, the Canadian government introduced the Safe Social Media Act, which would ban children under 16 from social media accounts and impose a statutory "Duty to Act Responsibly" on AI chatbot services. Chatbot providers would be required to mitigate the risk of harmful content, disclose their crisis-reporting thresholds (such as for self-harm conversations), and prevent harmful behaviour. Enforcement would fall to a new independent Digital Safety Commission.
Financial Stability Board issues 12-point AI governance guide for banks
The Financial Stability Board — the international body that coordinates financial rules across G20 economies — published a consultation report on 10 June setting out 12 sound practices for responsible AI adoption by financial institutions. The guidance covers organisation-wide AI governance and every stage of the AI lifecycle, from development to retirement. Boards and senior management are singled out as accountable for setting AI strategy and managing risk. The consultation remains open until 22 July 2026.