🔍 Read the full analysis: What Happens When European Governments Bring AI Into Public Work? 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
Developments reported across the Netherlands, Britain and Sweden show AI entering government work through staffing plans, asylum casework and rules on staff access. The evidence also includes a UK consultation flooded with responses identified as bots, while questions remain about AI accuracy, transparency and oversight.
European governments are taking sharply different approaches to AI in public work: the Netherlands is including it among tools for a planned reduction in central-government staffing, Britain is using generative AI to support asylum casework, and Sweden has blocked public versions of ChatGPT and Claude on government computers. A report published October 29 also describes errors in British asylum summaries and a UK health consultation in which most replies were flagged as bot-generated, raising questions about accuracy, accountability and public participation.
The Dutch government’s October 26 progress letter sets out cuts to central-government running costs that rise from €901 million in 2027 to €3.49 billion a year from 2035. It gives a rough estimate of about 25,000 fewer full-time positions, with police and defence exempt. The letter lists AI alongside other measures, including simpler rules, shared desks and standardised information technology; it does not present the job estimate as an AI-only reduction.
The same letter cautions against expecting too much from digitalisation and AI, saying earlier automation did not deliver the anticipated reductions in staffing and cost. It estimates that general efficiency can cover about a third of the planned cuts, with the rest requiring decisions about which government activities to stop. The cabinet says it plans to publish AI targets and savings plans in the first quarter of 2027.
In Britain, the Home Office uses two tools in asylum work: one summarises interview transcripts, and another searches official country information. The Open Rights Group (ORG), a digital rights organisation, said documents it obtained show the tools run on OpenAI models through Microsoft’s Azure cloud. The Home Office says trained caseworkers make decisions and that the tools do not decide asylum claims. Applicants are not specifically told that AI has been used in considering their claims, according to a statement by minister Anna Turley cited in the source report.
ORG reported that a 2026 Home Office spot check found 45 of 203 asylum summaries failed an initial review and 30 failed a second review. The pilot error rate had been 9%, the report said. Around 17,515 cases a month pass through the summary tool, while 203 summaries were checked in the cited review—about 0.3% of the estimated volume for April through July. Separately, Sweden’s Government Offices blocked browser versions of ChatGPT and Claude on staff computers from October 26, directing staff to approved tools instead.
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 Is Shaping Public Services
These cases show that bringing AI into government is not one policy choice. It can mean using software to support caseworkers, treating automation as one part of a wider cost-cutting programme, or restricting access to products that have not been approved for official work. Each approach affects public employees and the people who rely on government services.
The asylum example carries particular stakes because summaries can shape how officials review a person’s account, even if the software does not make the final decision. Reported errors and a small sample of checks make it harder to judge how reliably the tool works in live use. The lack of specific notice to applicants, as described by the minister, also leaves questions about how people can identify or challenge a faulty summary.
The health consultation episode points to a different risk: AI-generated or automated responses can distort a public process if they are counted as genuine submissions. The report says the UK health department flagged 90,835 of 118,756 replies to a 2023 consultation as bots and excluded them. That is 76% of the responses, according to the source. It says removing those replies changed the result, though the underlying consultation question and response breakdown are not detailed here.
Popular right now
De'Longhi Rivelia Automatic Espresso Machine with Milk Frother
Roborock F25 Ultra Steam Wet Dry Vacuum Mop
Ninja Hot & Iced XL Coffee Maker with Rapid Cold Brew, Black, CM371
As an affiliate, we earn on qualifying purchases.
Three Tests for Government AI
The Dutch staffing figure is tied to a broad budget and organisational programme, not a forecast of jobs that AI alone will eliminate. The October 26 letter estimates about 25,000 fewer full-time roles against a workforce of 160,016 full-time equivalents in the agencies and ministries covered at the end of 2025. Its warning that AI savings may be limited is part of the government’s own account of the plan.
The UK asylum tools were rolled out at different times: the policy-search tool reached decision-makers in July 2025, and the case-summary tool reached caseworkers in April 2026, according to the source. ORG published its report on October 12 after six months of parliamentary questions and freedom-of-information requests. The tools were not listed in the UK algorithmic transparency register at the time described in the report, despite the government’s AI Playbook asking departments to list such systems.
The source also says the EU AI Act treats AI used in asylum decisions and access to public services as high-risk. It reports that, after a “digital omnibus,” the relevant duties are due to begin on December 2, 2027, rather than in 2026. Sweden’s restriction, by contrast, concerns access to public versions of particular tools on staff computers; the report says approved alternatives remain available, so it is not a general ban on AI.
“Do not overestimate the staffing effects of digitalisation and AI.”
— Dutch government progress letter, sent to parliament by State Secretary Eric van der Burg on October 26
Accuracy and Oversight Questions
The available figures do not establish how often the asylum tools produce errors across all cases. The cited spot check covered 203 summaries, a small share of the estimated cases processed, and the source does not explain how those samples were selected or define every category of failure. It is also unclear whether applicants can access or correct summaries before a decision is made; the report says they cannot see or correct their own summary.
The source says the Independent Chief Inspector of Borders and Immigration asked ORG for more information while planning inspections for the following year. It does not say the inspector had confirmed a formal investigation. Nor does the supplied material establish what remedial steps, if any, the Home Office will take in response to the reported error rates or the tools’ absence from the transparency register.
For the Dutch plan, the exact number of jobs attributable to AI remains unknown: the government’s 25,000 estimate covers a wider cost-cutting programme, and AI savings plans are still pending. The consultation account also lacks detail here about the specific question, how bot responses were identified, and the revised totals after their removal. Those limits matter when interpreting claims about the effect of automated responses on the outcome.
Upcoming Plans and Reviews
The Dutch cabinet is expected to publish its AI targets and savings plans in the first quarter of 2027. Those documents should clarify where officials intend to apply AI and how much of the planned savings they expect it to deliver, though the October letter itself warns that past digitalisation did not reliably reduce staffing.
In Britain, the border inspector’s planned inspections next year may provide another opportunity to examine asylum casework tools. The inspector had requested more information from ORG, according to the source, but had not confirmed a probe. Further reporting or official disclosures may clarify the tools’ testing, monitoring, transparency-register status and the process available to applicants who dispute a summary.
The source places the relevant EU AI Act duties for high-risk uses on a December 2, 2027 start date. Until then, the immediate picture will depend on national rules and the way public bodies monitor their systems. Sweden’s approved-tool policy is already in effect, while the implications of the UK consultation’s bot-filtering and the Dutch staffing programme remain tied to further detail from the respective governments.
Key Questions
Is the Netherlands planning to cut 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 broad programme to reduce central-government costs. AI is one of several proposed tools, and the letter warns against overestimating its staffing effects.
Does AI make asylum decisions in the UK?
The Home Office says trained caseworkers make every decision and that the tools do not decide claims. The tools described support casework by summarising interview transcripts and searching official country information.
How many UK asylum summaries failed the cited check?
ORG said a Home Office spot check found that 45 of 203 summaries failed a first review and 30 failed a second. The source does not establish that this sample represents every summary processed.
Has Sweden banned AI in government?
No. The report says Sweden’s Government Offices blocked browser versions of ChatGPT and Claude on staff computers from October 26 and directed staff to approved tools. It describes a restriction on particular public versions, not a general ban on AI.
When do the relevant EU AI Act duties start?
The source says duties for the high-risk uses it describes are due to start on December 2, 2027, following a “digital omnibus.”
Source: ThorstenMeyerAI.com
