🔍 Read the full analysis: How AI Is Showing Up In European Government Systems on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Recent government actions in the Netherlands, Britain and Sweden show AI being considered for public-sector workforce changes, used in asylum casework and restricted on staff computers. A separate UK study found that bots accounted for most responses flagged in a 2023 health consultation. The cases raise questions about accuracy, transparency and public input, but they do not establish a single European policy or a common level of AI use.
Four developments reported this week show different ways AI is entering European government systems: the Netherlands is planning a smaller central-government workforce and lists AI as one possible efficiency tool; Britain is using generative AI to support asylum casework; Sweden has blocked public versions of ChatGPT and Claude on Government Offices computers; and a UK health consultation review found most responses it flagged came from bots. The cases matter because they put staffing, decisions affecting applicants, and the reliability of public consultation under scrutiny.
The Netherlands’ October 26 progress letter to parliament sets out cuts to central-government running costs that rise to €3.49 billion a year from 2035. The letter describes the projected staffing impact as a rough estimate of about 25,000 full-time positions. It lists AI alongside other measures, including simpler rules, standardised IT and shared desks. The cabinet plans to publish AI targets and savings plans in the first quarter of 2027. The letter also cautions against overestimating AI’s staffing effects, saying earlier digitalisation and automation did not deliver the promised staff and cost savings.
In Britain, the Home Office uses two tools in asylum casework, according to documents obtained by the digital rights group Open Rights Group (ORG). One summarises asylum interview transcripts; another searches official country information. ORG says both use OpenAI models through Microsoft’s Azure cloud. A Home Office spot check found that 15% to 22% of sampled summaries failed review, compared with a 9% inaccuracy or incompleteness rate in the earlier pilot. ORG says only about 0.3% of summaries were checked. The Home Office says trained caseworkers make decisions and that the tools do not decide claims.
Sweden’s Government Offices blocked browser versions of ChatGPT and Claude on staff computers from October 26, according to the source report. Staff are directed to approved tools; the measure is described as a restriction on contracts and data use, not a general ban on AI. Separately, a study of a 2023 UK health consultation found that the department flagged 90,835 of 118,756 responses as bots and excluded them. The report says removing those replies changed the consultation’s result, but the source material does not provide the underlying study methods or the competing response totals.
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
AI Touches Staffing and Casework
Together, the cases show that government AI is not one policy or one kind of system. It can be part of a long-term workforce plan, a support tool used while officials process asylum claims, a technology barred from work devices, or a factor in deciding which public submissions count. Each use has different consequences and needs different oversight.
The asylum example carries immediate stakes for people whose accounts are summarised. If a summary omits or changes information, that could affect how a caseworker reads a claim, even if the software does not formally make the decision. The reported failure rates and limited checking make human review and applicants’ ability to correct records central questions. In the Netherlands, the letter’s warning matters because it separates an actual budget plan from the still-uncertain contribution AI might make to reducing staffing.
The consultation case points to a different risk: automated responses may distort a public process, but methods used to detect bots can also determine which views are counted. Readers and policymakers need to know how responses were classified and what changed after exclusion before drawing conclusions about public opinion.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Different Rules, Different Uses
The Dutch plan is framed as a broader government efficiency and spending programme, not an AI-only jobs initiative. The October 26 letter projects annual cuts rising from €901 million in 2027 to €3.49 billion from 2035. It says the 25,000-position estimate is rough, and that general efficiency may cover only about one-third of the planned reduction; the rest would require choices about what government stops doing. Police and defence are exempt from the specified cuts.
The British Home Office tools have been rolled out in stages: its policy-search tool reached decision-makers in July 2025, while the transcript-summarising tool reached caseworkers in April 2026, according to ORG. ORG published its findings on October 12 after parliamentary questions and freedom-of-information requests. It also says applicants are not specifically told that AI has been used and cannot see or correct the generated summary. The source report says neither tool appears in the UK’s algorithmic transparency register.
The UK consultation received 118,756 responses in 2023, with 76% flagged as bots by the health department, according to the study described in the source. The EU AI Act is also relevant to public-sector systems: the source says AI used in asylum decisions and access to public services is treated as high-risk, with the relevant duties scheduled to begin on December 2, 2027 after this year’s digital omnibus changes.
“Do not overestimate the staffing effects of digitalisation and AI.”
— The Dutch government’s October 26 progress letter to parliament
Accuracy and Oversight Gaps
Several details remain unsettled. The Dutch figure is an estimate of positions associated with a wider spending plan, not a confirmed number of jobs that AI will eliminate. The government has not yet published its planned AI targets and savings plans; those are expected in the first quarter of 2027.
For the UK asylum tools, the source material does not establish how often summaries contain consequential errors or whether a failed review changed an outcome. The Home Office spot check covered 203 summaries, while ORG’s estimate of monthly use suggests a much larger volume. The department’s account of human decision-making does not resolve questions about disclosure, access to summaries, or how errors are corrected. The source also says the Independent Chief Inspector of Borders and Immigration requested more information from ORG while planning next year’s inspections, but did not confirm a formal probe.
For the consultation, the supplied account does not explain the bot-detection method, its error rate, or the figures behind the reported change in result. Nor does the Sweden account specify which approved tools staff may use or how the restrictions will be reviewed. These limits mean the cases should not be treated as evidence of a uniform European approach.
Targets, Inspections and Rules
The next stated Dutch milestone is the publication of AI targets and savings plans in the first quarter of 2027. Those plans should clarify what role the government expects AI to play alongside spending cuts and other administrative changes.
In Britain, the Home Office’s use of the asylum tools may receive further scrutiny through the border inspector’s planned inspections next year. The inspector has asked ORG for information, but whether that leads to a specific inquiry remains unconfirmed. Further reporting or official disclosures may also clarify the scale of checking, whether applicants will be informed, and how they can challenge errors.
Across the EU, the source places the relevant AI Act duties for high-risk uses on a schedule beginning December 2, 2027. In the meantime, the reported cases leave practical questions for governments: how to measure accuracy, disclose AI use, protect sensitive data, and distinguish genuine public participation from automated submissions.
Key Questions
Is the Netherlands cutting 25,000 jobs because of AI?
Not according to the reported letter. The figure is a rough estimate linked to a broader central-government spending-cut plan. AI is one of several possible efficiency tools, and the letter 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. The tools are used to summarise interview transcripts and search official country information, according to documents obtained by ORG.
How often were the asylum summaries checked?
ORG says experts reviewed 203 summaries in a spot check and estimates that this was about 0.3% of the summaries produced over the period it examined. The reported check found that 15% to 22% failed review, depending on the review stage.
Has Sweden banned government use of AI?
No. The reported restriction blocks public browser versions of ChatGPT and Claude on Government Offices computers. Staff are directed to approved tools; the measure is not described as a general ban on AI.
When do the EU AI Act duties cited in the report begin?
The source says the relevant duties for high-risk AI uses in areas such as asylum decisions and access to public services are scheduled to begin on December 2, 2027, following changes under the digital omnibus.
Source: ThorstenMeyerAI.com
