🔍 Read the full analysis: Could AI Reshape Public Services Across Europe? on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Developments reported on October 29 show European governments taking sharply different approaches to AI in the state: the Netherlands includes AI among tools for a planned civil-service reduction, the UK uses AI in asylum casework and faces scrutiny over accuracy and disclosure, and Sweden has blocked public ChatGPT and Claude on government computers. A separate UK consultation was heavily affected by bot responses. The examples show what is known about use and safeguards, but do not establish that AI will deliver large staff savings or that the UK tools have caused specific asylum decisions.
European governments are taking different steps to bring artificial intelligence into public services, with recent developments involving staffing plans in the Netherlands, asylum casework in Britain and restrictions on public AI tools in Sweden. The cases put questions about efficiency, accuracy and public accountability into focus, but do not show that AI alone will deliver government job cuts or make asylum decisions.
In the Netherlands, a government progress letter sent to parliament on October 26 estimates that planned reductions in central-government running costs could translate into about 25,000 fewer full-time positions by 2035. The figure is described in the letter as a rough estimate. The annual savings target rises from €901 million in 2027 to €3.49 billion from 2035. AI is listed alongside other measures, including simpler rules, shared desks, standardised IT and joint purchasing. The letter also warns against overestimating how much staff digitalisation and AI can save.
In Britain, the Home Office uses two generative AI tools in asylum work: one to summarise interview transcripts and another to search official country information. The Open Rights Group, a digital rights organisation, reported that people seeking asylum are not specifically told when AI has been used and cannot see or correct the summaries. In a Home Office check of 203 summaries in 2026, 45 failed the first review and 30 failed a second review. The report says only about 0.3% of summaries were checked, based on the volume of cases described. The Home Office says trained caseworkers make every decision and the tools do not decide claims.
Two other developments raise related governance questions. A UK health consultation in 2023 received 118,756 responses; the health department identified 90,835, or 76%, as bot submissions and excluded them. In Sweden, the Government Offices blocked browser versions of ChatGPT and Claude on staff computers from October 26, directing staff to approved tools instead. The reported restriction concerns public versions of those services, not a ban on using AI.
Infographic 01 · The Netherlands
25,000 fewer civil servants. AI is one tool, not the reason.
The driver is a budget cut. The government plans to use AI, and warns against expecting too much from it.
The cut grows every year
Cuts to the running costs of central government, € billion per year.
Police and defence are exempt. Many organisations face about 20% fewer staff, some up to 45%.
Where the money must come from
The letter’s own split of the cut.
AI is one of several tools inside both parts, next to simpler rules, standard IT, shared desks and joint purchasing.
AI in the plan
- “Ambitious targets and savings plans” for AI in Q1 2027
- Purchasing to examine “systematically” where AI fits
- Conditions: security, data quality, sovereignty
Also in the plan
- At least 500 rules a year scrapped or simplified
- One desk per two employees
- Fewer advisory bodies (28 councils, ~150 committees)
The letter’s own warning
- Past digitisation did not deliver the promised staff cuts
- It brought new tasks and more contractors
- Cuts may affect service and continuity
“Do not overestimate the staffing effects of digitalisation and AI.”Lesson 2 in the Dutch government’s own letter, 26 Oct 2026 · translation: Thorsten Meyer
Sources: BZK, Voortgangsbrief Slagvaardige Overheid + annex, 26 Oct 2026; Rijksoverheid news release, 26 Oct 2026. 25,000 ÷ 160,016 ≈ 1 in 6 and bar scaling: Thorsten Meyer.
Powered by Thorsten Meyer AI
Staffing, Asylum and Public Trust
Together, the cases show that public-sector AI is not a single policy choice. Governments are considering it as a way to handle administrative work, using it to support decisions that affect people, and setting limits on which services staff may access. Each use raises distinct questions about who checks the output, whether affected people can challenge errors, and how public bodies explain their use of the technology.
The Dutch figures matter because they connect AI with a substantial planned reduction in government staffing. But the cabinet’s letter does not attribute the 25,000 estimate to AI alone: it lists several methods and says efficiency may cover only about one-third of the spending cut. The remaining savings require decisions about which government activities to reduce or stop. Treating the plan as a straightforward AI-driven job-cut programme would go beyond the evidence provided.
The UK asylum case is more immediate for individuals. A summary of an interview about why someone fled could affect how a caseworker understands the evidence, even if the tool does not make the formal decision. The reported discrepancy between the pilot and later checks, alongside limited checking and a lack of notice, makes transparency and review central concerns. The consultation figures also show that automated submissions can affect how public feedback is counted, though they do not by themselves establish how the department’s final policy was shaped.
Products Worth Considering
Logitech MX Keys S Wireless Keyboard, Graphite
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
Apple MacBook Pro 14‑inch (2021) with M1 Pro, 16 GB RAM, 512 GB SSD – Space Gray
As an affiliate, we earn on qualifying purchases.
Three Governments, Different Rules
The Dutch progress letter sets out a multi-year programme called a “decisive government,” combining budget reductions with administrative changes. The organisations covered employed 160,016 full-time equivalents at the end of 2025, according to the letter; police and defence are exempt. It says AI savings targets and plans are expected in the first quarter of 2027. The same document recalls that earlier digitisation and automation did not produce the promised staff and cost savings, and sometimes created new work and reliance on contractors.
The Open Rights Group published its findings on October 12, after six months of parliamentary questions and freedom-of-information requests. According to documents it obtained, both Home Office tools run on OpenAI models through Microsoft’s Azure cloud. The search tool reached all decision-makers in July 2025, while the summarisation tool reached all caseworkers in April 2026. The group’s account says a 2025 pilot found 9% of summaries inaccurate or incomplete; its later account of the Home Office check reports 15% failing a second review and 22% failing the first.
These developments sit alongside EU rules: the source report says the AI Act classifies AI used in asylum decisions and access to public services as high-risk, with relevant duties due to begin on December 2, 2027 after changes described as the “digital omnibus.” That timing provides a regulatory backdrop, but it does not settle the specific concerns raised about the UK tools or the national choices in the Netherlands and Sweden.
“Do not overestimate the staffing effects of digitalisation and AI.”
— The Dutch government’s October 26 progress letter to parliament
Accuracy, Oversight and Savings
The source material does not establish how many Dutch jobs, if any, will be removed specifically because of AI. The government has not yet published the promised AI targets and savings plans, and its estimate of 25,000 fewer positions is explicitly rough. The split between savings from efficiency and cuts to services will depend on later decisions.
For the UK asylum tools, the reported checks do not establish the error rate across every summary or whether errors changed case outcomes. The small sample of 203 and the reported limited checking leave uncertainty about the scale and effect of inaccuracies. The source says the Independent Chief Inspector of Borders and Immigration asked the Open Rights Group for more information while planning next year’s inspections; it does not say the inspector confirmed a formal investigation. The tools’ absence from the UK algorithmic transparency register is also reported, but the material does not explain whether or when they will be added.
It is also not clear how the health department identified bot responses, whether any genuine submissions were misclassified, or how the excluded responses affected the consultation’s policy outcome. Sweden’s restrictions are described as a decision about contracts and data access; further details on the approved tools and how their use is monitored are not provided.
Plans, Inspections and AI Rules
The Dutch government says it will publish AI targets and savings plans in the first quarter of 2027. Those plans should provide a clearer account of where officials expect AI to be used and how much of the cost-reduction programme, if any, is assigned to it.
In Britain, the next reported milestone is the Independent Chief Inspector of Borders and Immigration’s inspections planned for next year. The inspector has requested information from the Open Rights Group, but has not confirmed a probe, according to the source report. Further information from the Home Office about accuracy checks, applicant notice and transparency-register entries would clarify how the tools are governed.
The source report gives December 2, 2027 as the start date for relevant EU AI Act duties following the digital omnibus changes. That date is a future regulatory milestone, not a resolution of the current questions. For now, the confirmed picture is of governments testing different approaches while the evidence on accuracy, savings and accountability remains incomplete.
Key Questions
Is the Netherlands cutting 25,000 jobs because of AI?
No. The government letter gives a rough estimate of about 25,000 fewer full-time positions by 2035 as a result of wider cost reductions. It lists AI as one tool among several and warns against overestimating its staffing effects.
Does AI decide asylum claims in Britain?
The Home Office says trained caseworkers make every decision and that its tools do not decide claims. One tool summarises interview transcripts and another searches country information, according to the source report.
What did the UK asylum-tool checks find?
The Open Rights Group reported that a 2026 Home Office spot check of 203 summaries found 45 failed the first review and 30 failed a second review. The figures do not establish the error rate across all summaries or whether errors changed case outcomes.
Has Sweden banned AI in government?
No. The reported action blocked browser versions of ChatGPT and Claude on Government Offices computers from October 26. Staff were directed to approved tools; the measure is described as a restriction on those services, not a general AI ban.
When do the relevant EU AI Act duties begin?
The source report gives December 2, 2027 as the start date for relevant duties after changes described as the digital omnibus. That timetable does not resolve the current questions about individual national systems.
Source: ThorstenMeyerAI.com
