Three posts from the past two days that bear directly on how you work with AI at work. The cabinet is setting its own AI course, the NCTV sees attackers using AI at every stage of a cyberattack and Microsoft is limiting free Copilot Chat for work accounts.
Cabinet vision: "Choosing AI for everyone"
On 6 November, the Council of Ministers adopted the cabinet vision on AI, on the proposal of State Secretary Willemijn Aerdts (Digital Economy and Sovereignty). The vision has three priorities: responsible development and application, access to advanced AI and safety, and an AI transition guided by public values. Concretely, the cabinet wants to remove barriers for SMEs, procure AI applications for government jointly and create room for experimentation under supervision. The cabinet is also exploring a Dutch AI safety institute. An elaboration of AI infrastructure and computing capacity follows at the beginning of 2027. Healthcare, energy, food production, security and the high-tech industry are the sectors the cabinet is focusing on.
What does this mean for you? If you work in government, you can expect joint procurement in future instead of separate choices per organisation. No figure for budget or planning per measure is given in the report yet. Until then, your own assessment remains necessary: the GDPR check for Copilot in government shows how a widely used model scores. In healthcare and industry, it remains to be seen what the experimentation room will precisely entail.
NCTV: AI makes attacks faster and bigger
According to the Cybersecuritybeeld Nederland 2026, also from 6 November, malicious actors demonstrably use AI in all phases of cyberattacks. They develop malware with it, write more convincing phishing emails and personalise disinformation on a large scale. Small organisations and SMEs, according to the NCTV, continue to lag behind on measures for their digital resilience. Without significant improvement, the NCTV expects the number of incidents to rise sharply in the coming years.
What does this mean for you? The advice is basic but concrete: patch management, a small attack surface, logging and monitoring, and including AI risks in your security policy. That also includes which AI tools your colleagues use with company data. If you work in cybersecurity, the guide to cybersecurity shows which tools exist, and the GDPR check for Copilot in cybersecurity shows what happens to your input.
Copilot Chat: limits from 9 November
According to message MC1536163 in the Microsoft 365 Message Center, Microsoft is introducing usage limits for Copilot Chat for users without a Microsoft 365 Copilot licence in business environments. This concerns Copilot Chat in the apps on web, desktop and mobile; the rollout begins on 9 November. Microsoft does not state any numbers. Administrators cannot view or adjust the limits. If you reach the limit, you receive a notification in the app and can try again later; an answer already in progress is not interrupted.
What does this mean for you? If your team relies on the free Copilot Chat, inform your helpdesk and check whether a licence is needed. If you want to weigh alternatives, compare ChatGPT and Copilot or Claude and Copilot. What Copilot does with your data is set out on the model page for Copilot.
In brief
- The cabinet wants joint AI procurement, room for experimentation, and is exploring an AI safety institute; computing power follows early 2027.
- The NCTV sees AI at every stage of cyberattacks and asks you to include AI risks in your security policy.
- Free Copilot Chat on work accounts will get limits from 9 November, without fixed numbers.
- All models and privacy scores are at models and safety.