4 November 2026 · Who decides how AI is used

This morning’s selection brings together three documents published on 3 November: Google’s education updates, the British warning on cybersecurity and António Costa’s letter preparing the European debate on AI. These are developments from the previous day, covered in the morning briefing: available features, announced changes and risk scenarios have different timelines and meanings.

Google · Guided Learning comes to Classroom

Google announced the global availability of Guided Learning in Classroom: teacher-assigned activities that guide students through reasoning using class materials. For Search, however, the programme will begin with a pilot in selected US schools during November. The company also announced controls that will keep Gemini, Gemini Notebook, Gemini in Classroom and Google Search switched off for K-12 student accounts until administrators enable them. Google will share details before introducing the changes. Availability in Classroom and the future change to controls should not be confused.

Google · Guided Learning and school controls, 3 November 2026 ↗

United Kingdom · Cyber risk grows through existing weaknesses

The National Cyber Security Centre considers it highly likely that, by 2028, a ransomware attack against British critical infrastructure will involve AI-powered malware. This is the agency’s forward-looking assessment, not a description of an attack that has already happened. The report focuses on neglected vulnerabilities, insufficient investment and systems unnecessarily exposed to the Internet: AI can accelerate the discovery and exploitation of these weaknesses. Its message to organisations is to strengthen basic defences and operational continuity now, also considering the use of AI for protection.

NCSC · AI and critical infrastructure, 3 November 2026 ↗

Europe · Dublin to discuss autonomy and safety

In the letter released on 3 November, António Costa announces that the informal meeting of European leaders in Dublin on 13 November will focus on artificial intelligence. The programme includes a discussion with Arthur Mensch, CEO of Mistral AI, on competitiveness, infrastructure and reducing dependencies. Geoffrey Hinton will address social effects, safety and trust. The summit has yet to take place: the letter sets its agenda without announcing decisions already adopted. For Europe, industrial capability and the protection of people are thus part of the same political discussion.

European Council · António Costa’s letter on the 13 November meeting, 3 November 2026 ↗

The desk’s view · Control becomes a practical choice

Schools decide which tools to enable; businesses which vulnerabilities to fix; institutions which capabilities to build. The desk’s view is that adopting AI requires verifiable choices alongside announcements. For readers, the useful distinction remains simple: what is available today, what is planned and what is a forecast.

Sources and documents

Document date or reference period: Selection for 4 November 2026 · Documents published on 3 November 2026.

Desk selection and summary with AI assistance, based on the linked original documents. Available features, future announcements and risk assessments are distinguished; the conclusion is the desk’s interpretation.

Report an update or correction →