🔍 Read the full analysis: 25,000 Jobs, Asylum Files And Bot Replies: Europe’s AI Push on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A report dated October 29 describes four distinct moves in European government AI: the Netherlands plans about 25,000 fewer central-government jobs by 2035, Britain uses AI tools to summarize asylum interviews, a UK health consultation received tens of thousands of replies classified as bots, and Sweden blocked public ChatGPT and Claude on staff computers. The cases raise questions about workforce cuts, accuracy and public input, but do not establish that AI alone will cause the Dutch reductions or make asylum decisions.
European governments are taking sharply different approaches to artificial intelligence, according to a report published October 29: the Netherlands is planning about 25,000 fewer central-government jobs by 2035, Britain is using AI to summarize asylum interviews, and Sweden has blocked public versions of ChatGPT and Claude on government staff computers. The cases also include a UK health consultation in which officials classified 76% of responses as bot-generated, highlighting concerns about both AI in public administration and AI-generated input to government.
The Dutch figure comes from a progress letter sent to parliament on October 26 by State Secretary Eric van der Burg. It estimates around 25,000 full-time positions could be removed as central-government running-cost reductions rise to €3.49 billion annually from 2035. The letter describes the job estimate as rough. AI is listed alongside other measures, including simpler rules, shared desks, standardised IT and joint purchasing; the plan is not solely an AI-led jobs program. Police and defence are exempt from the cuts.
In Britain, the Home Office uses two generative AI tools in asylum casework: one summarizes interview transcripts, while another searches official country information. Documents obtained by the Open Rights Group (ORG) indicate both use OpenAI’s GPT-4o through Microsoft’s Azure cloud. The Home Office says trained caseworkers make every decision and that the tools do not decide asylum claims. ORG reports that applicants are not specifically told AI was used and cannot see or correct the generated interview summary.
ORG’s review found that 15% to 22% of 203 summaries examined in a 2026 spot check failed review, compared with a 9% error rate in the 2025 pilot. The report says about 17,515 cases a month pass through the summary tool, while only 203 summaries were checked. Separately, the UK health department classified 90,835 of 118,756 replies to a 2023 consultation as bots and excluded them. The source report says the result changed when those replies were removed, but gives no further details here about the consultation’s final policy outcome.
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
Pressure on Public Services and Decisions
Together, the cases show that government AI is not one policy choice. It can be presented as a way to reduce costs, used to assist staff handling high-stakes cases, or restricted because of data and procurement concerns. The consequences differ: a staffing target affects public-sector capacity, while an inaccurate asylum summary could shape how a caseworker understands an applicant’s account, even if a person retains formal decision-making authority.
The figures also point to a gap between scale and oversight. ORG’s reported sample of 203 summaries is small against monthly use of about 17,515 cases. In public consultation, automated submissions can distort the apparent weight of opinion. That makes transparent procedures, review mechanisms and clear disclosure relevant to both people affected by government decisions and the public trying to influence policy.
The Dutch letter cautions against treating AI as a ready-made substitute for staff. It says earlier digitisation and automation did not deliver the promised staffing and cost savings, and estimates general efficiency gains would account for only about one-third of the planned reductions. The remaining cuts would require decisions about which government activities to stop.
Products Worth Considering
EcoFlow DELTA 3 Classic Portable Power Station, 1024Wh, 1800W
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
DJI Osmo Pocket 3 Vlogging Camera with 1-Inch CMOS Sensor and 4K Video
As an affiliate, we earn on qualifying purchases.
Four Government Decisions Behind the Report
The Dutch central-government organisations covered by the cuts employed 160,016 full-time equivalents on December 31, 2025, according to the letter. That was 41% more than in 2018. The proposed savings rise from €901 million in 2027 to €3.49 billion a year from 2035. The cabinet plans to publish AI targets and savings plans in the first quarter of 2027, while the letter also warns officials not to overstate AI’s likely effect on staffing.
ORG published its findings on the Home Office tools on October 12 after six months of parliamentary questions and freedom-of-information requests. The tools reached decision-makers and caseworkers in 2025 and 2026, respectively. On October 27, WIRED reported that the Independent Chief Inspector of Borders and Immigration had asked ORG for more information while planning inspections for the following year; that did not amount to confirmation of a formal probe.
In Sweden, the Government Offices blocked browser access to public ChatGPT and Claude from October 26. The report characterises this as a restriction tied to contracts and data handling, not a ban on AI: staff are directed to approved tools. It also notes that the EU AI Act treats systems used in asylum decisions and access to public services as high-risk, with relevant duties scheduled to begin on December 2, 2027 after changes described as the “digital omnibus.”
Accuracy, Oversight and Outcomes
The available information does not establish how often errors in the asylum summaries changed a case outcome, or whether applicants can challenge or correct a summary through another route. The Home Office’s reported spot check covered 203 summaries, and the source does not provide details about how those cases were selected. The department’s position that staff decide claims does not by itself show how they use, verify or weigh AI-generated summaries.
It is also unclear whether the immigration inspector will open a formal investigation: the report says the inspector requested information while preparing future inspections, but did not confirm a probe. For the Dutch plan, 25,000 is a rough estimate attached to a broader budget-cutting program, not a confirmed count of jobs that AI will eliminate. The source gives no detailed breakdown of which roles or services would be affected.
The bot classification in the health consultation is reported by the department, but the source does not explain the detection method, how classifications were challenged, or the full policy consequences of excluding the responses. The scale of automated submissions is clear from the reported count; their origin and the reliability of each individual classification are not established in the material.
Plans and Reviews Due in 2027
The Dutch cabinet is expected to publish its AI targets and savings plans in the first quarter of 2027. Those documents should clarify what role officials expect AI to play alongside other spending and organisational changes. The broader savings schedule extends to 2035.
In Britain, the Home Office’s use of the asylum tools may receive further scrutiny through the Independent Chief Inspector’s planned inspections, though no investigation has been confirmed. The reported accuracy figures and limited sampling leave questions about how frequently summaries are checked and what protections applicants can access.
The UK’s high-risk AI duties under the EU AI Act are scheduled, according to the source report, to take effect on December 2, 2027. Until then, the immediate developments to watch are the Dutch plans, any further disclosures or scrutiny of the Home Office tools, and how public bodies document and review AI use.
Key Questions
Is the Netherlands cutting 25,000 jobs because of AI?
Not according to the report. The estimate is part of a wider plan to reduce central-government costs. AI is one proposed tool among several, and the Dutch letter warns against overestimating its staffing effects.
Does AI decide UK asylum claims?
The Home Office says trained caseworkers make every decision and that its tools do not decide claims. One tool summarizes interview transcripts; ORG says applicants are not specifically told when AI is used and cannot review or correct the summary.
How many UK asylum summaries failed the reported check?
In a 2026 spot check of 203 summaries, 45 failed the first review, or 22%, and 30 failed the second, or 15%, according to ORG. The source does not explain whether the two review figures represent separate samples.
Why were most replies to the 2023 health consultation excluded?
The health department classified 90,835 of 118,756 replies as bot-generated and excluded them, according to the report. The material provided does not describe the detection method or the department’s process for reviewing individual classifications.
Did Sweden ban AI in government?
No. The reported action blocks staff access to the public browser versions of ChatGPT and Claude. Government staff are directed to approved tools; the report describes the restriction as related to contracts and data, not a general ban on AI.
Source: ThorstenMeyerAI.com
