🔍 Read the full analysis: The Public Tasks Where Europe’s Governments Are Using AI on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Recent developments in the Netherlands, the UK and Sweden show how AI is entering government work: as a possible efficiency tool, a support tool in asylum casework and a source of data-governance concerns. The source report also says bots made up 76% of replies to a 2023 UK health consultation. The scale of AI’s role in Dutch job reductions and the full effect of errors in UK asylum summaries remain unclear.
European governments are taking sharply different approaches to artificial intelligence in public services, with the Netherlands linking AI to a plan for about 25,000 fewer central-government jobs, the UK using AI-generated summaries in asylum casework, and Sweden blocking public versions of ChatGPT and Claude on staff computers. The examples, reported by ThorstenMeyerAI.com on October 29, also include a UK health consultation in which the government classified 76% of replies as bot submissions.
The Dutch government’s October 26 progress letter sets out rising cuts to central-government running costs, reaching €3.49 billion annually from 2035. It describes the estimate of around 25,000 full-time jobs as rough. AI is listed among several ways to make government work more efficient, alongside simpler rules, shared desks, standardised information technology and joint purchasing. The letter also cautions against overestimating AI’s staffing effects, citing past digitisation efforts that did not deliver promised savings.
In Britain, the Home Office uses two tools in asylum casework: one summarises interviews with applicants, and another searches official country information. Documents obtained by the Open Rights Group say both use OpenAI models through Microsoft’s Azure cloud. The government says trained caseworkers make asylum decisions and the tools do not decide claims. However, according to the group’s report, applicants are not specifically told when AI has been used and cannot see or correct the summaries of their interviews.
The report gives a Home Office pilot error rate of 9% of summaries. In a later spot check of 203 summaries, Home Office experts found 45 failed an initial review and 30 failed a second review, equivalent to 22% and 15% of the sample, respectively. The source says about 17,515 cases a month go through the summary tool, while 203 summaries were checked—about 0.3% of the estimated April-to-July volume. Separately, the UK health department flagged 90,835 of 118,756 responses to a 2023 consultation as bots and excluded them.
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’s Role in Public Decisions
These cases put different public responsibilities in view. A tool that summarises an asylum interview may save staff time, but errors can matter when officials are handling a person’s account of why they fled. The Home Office says human caseworkers retain decision-making responsibility; the reported error rates and limited sample checking raise questions about how reliably staff can spot omissions or mistakes.
The Dutch plan illustrates another distinction: AI is one proposed efficiency measure, not the stated cause of all job reductions. The government’s own letter says general efficiency may cover only about one-third of the planned savings, with the remainder requiring decisions about which activities to stop. In the UK consultation, excluding responses classified as bots changed the result, according to the report, making the integrity of public input relevant to policy outcomes.
Sweden’s restriction highlights a separate issue: whether staff may use widely available AI services with government information. The measure reported is a block on public browser versions, not a general prohibition on AI. Together, the examples show that governments are making choices about staffing, decision support, participation and data handling—not simply adopting or rejecting the technology.
Products Worth Considering
Dell XPS 13 9350 AI Professional Laptop 13.4 120Hz FHD+ Display
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite
As an affiliate, we earn on qualifying purchases.
Four Different Government Choices
The Dutch estimate sits within a broader cost-reduction programme. The October 26 letter says cuts rise from €901 million in 2027 to €3.49 billion a year from 2035. The covered ministries, inspectorates and agencies employed 160,016 full-time equivalents at the end of 2025, according to the letter; police and defence are exempt. The cabinet plans to publish AI targets and savings plans in the first quarter of 2027.
In the UK, the Asylum Policy Search tool was rolled out to decision-makers in July 2025, while the Asylum Case Summarisation tool reached caseworkers in April 2026, according to the source report. The Open Rights Group published its findings on October 12 after six months of parliamentary questions and freedom-of-information requests. The report says neither tool appears in the government’s algorithmic transparency register, despite guidance asking departments to list such systems.
Sweden’s Government Offices blocked public versions of ChatGPT and Claude on staff computers from October 26. Staff are directed to approved tools. The source characterises this as a decision about contracts and data, rather than a ban on AI. The report also notes that EU rules classify AI used in asylum decisions and access to public services as high-risk, with relevant duties scheduled to begin on December 2, 2027, following changes described as the digital omnibus.
“Do not overestimate the staffing effects of digitalisation and AI.”
— The Netherlands’ government progress letter, as quoted by ThorstenMeyerAI.com
Oversight and Error Rates
It is not clear how often the asylum tools’ outputs are reviewed in routine use beyond the cited checks, or how errors affect individual decisions. The reported 203-summary review is a small sample relative to the estimated volume, and the source does not provide details on how the sample was selected. The Home Office’s position is that caseworkers decide claims, but the report says applicants cannot inspect or correct the summaries.
The Dutch estimate is explicitly described as rough, and the eventual share of job reductions attributable to AI has not been specified. The planned targets and savings plans are due in the first quarter of 2027. For the UK health consultation, the source reports the number of responses flagged and says the result changed when those responses were excluded, but the supplied material does not detail the method used to identify bots or the substantive outcome of the consultation.
Sweden’s policy is described as applying to public versions of two named chatbots on staff computers. The source does not give the full terms of the approved-tool arrangements or say whether the block will change. The report says a UK border inspector sought more information about asylum AI ahead of future inspections, but does not confirm that a formal investigation has begun.
Targets, Inspections and Rules
The next stated Dutch milestone is publication of the cabinet’s AI targets and savings plans in the first quarter of 2027. Those plans may clarify where officials expect AI to contribute and how the government will distinguish technology-related savings from broader cuts in public activity.
In the UK, the Independent Chief Inspector of Borders and Immigration has asked the Open Rights Group for more information while planning inspections for next year, according to a report by WIRED cited in the source. That request does not confirm a formal probe. Further scrutiny could clarify how the Home Office tests summaries, records the tools’ use and handles errors, but no specific inspection findings are yet available.
European governments will also face the scheduled start of relevant EU AI Act duties on December 2, 2027, as described in the source material. Until then, the immediate questions are practical: whether public bodies disclose AI use, how they check its outputs, and what limits they set on staff use of external services. The reported examples do not establish a single European policy or show that AI has replaced public-sector decision-makers.
Key Questions
Is the Netherlands cutting 25,000 jobs because of AI?
No. The Dutch letter gives a rough estimate of about 25,000 fewer full-time positions as part of a wider cost-reduction programme. It lists AI as one possible efficiency tool and warns that its staffing effects should not be overestimated.
Does AI decide UK asylum applications?
The Home Office says trained caseworkers make every decision and that its tools do not decide claims. One tool summarises interview transcripts; another searches official country information. The Open Rights Group’s report raises concerns about disclosure, checking and summary accuracy.
How often were the UK asylum summaries checked?
The source report says experts reviewed 203 summaries against an estimated 70,000 cases processed from April to July, or about 0.3%. It does not establish that this was the only checking carried out in routine operations.
Did Sweden ban government use of AI?
No general ban is reported. From October 26, Sweden’s Government Offices blocked public browser versions of ChatGPT and Claude on staff computers, while directing staff to approved tools.
When do the cited EU AI Act duties begin?
The source says duties for the relevant high-risk uses are scheduled to begin on December 2, 2027, after changes described as the digital omnibus. The article does not set out the full scope of those requirements.
Source: ThorstenMeyerAI.com
