Why Britain’s Bureaucracy Is Seeing More AI-Driven Paperwork

🔍 Read the full analysis: Why Britain’s Bureaucracy Is Seeing More AI-Driven Paperwork on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Complaints, Freedom of Information requests and employment tribunal claims have risen across several UK services since ChatGPT launched in November 2022, according to data examined in a Sky News report published October 20. The pattern raises questions for public bodies facing more written submissions, but the timing alone does not establish that AI caused the increases or that the new submissions are invalid.

Sky News technology correspondent Rowland Manthorpe reported increases across nine types of written complaint or claim. Household complaints about water companies rose 144%; complaints about English local authorities increased 78%, written complaints to GP practices 71%, and complaints to the Department for Work and Pensions (DWP) 64%. Individual employment tribunal claims rose 62%, while complaints about legal services were up 49%. The report gives these as growth since ChatGPT launched; its cited figures do not, in the supplied account, specify the precise measurement windows or underlying totals for each category.

The report says the trend appears across channels with different circumstances, including water services, councils, health care and benefits. It describes a broadly similar pattern: volumes were relatively flat for years, began rising around late 2022 and then accelerated. Each channel has its own possible drivers. For example, water companies have faced difficult years, so the parallel trend does not show that AI is the sole or direct cause.

Manthorpe also tested 100 recent Freedom of Information requests with an AI-text detector. The tool classified 89% as fully AI-written and 9% as human-written; a smaller 2019 sample was classified as entirely human. Sky News cautioned that the test is not conclusive: AI detectors can be unreliable, and a sample of requests cannot prove AI caused the wider increase.

More Submissions, More Work for Officials

The increase matters because public services rely heavily on written processes. Complaints, information requests and appeals can require staff to review evidence, check rules and respond within formal procedures. If AI makes it easier and faster for people to prepare submissions, officials may face a larger workload, including requests that are lengthy, unclear or inaccurate. A survey of civil servants cited by Sky News found that 58% said AI-generated material from the public significantly affects their work.

Higher volumes do not automatically mean a wave of frivolous claims. Sky News reported that the proportion of DWP complaints upheld had fallen from the 50% range to close to 40%, but remained substantial. Since the total number of complaints has grown, a lower success rate does not by itself show that the number of upheld complaints has fallen. Researcher Chris Schmitz, whose work on possible cases of “agentic flooding” was reported by TechCrunch, said most cases he examined involved people claiming something to which they were entitled.

That distinction matters for policy. Adding fees or barriers could deter legitimate claims alongside poorly formed submissions. At the same time, public bodies need workable ways to handle increased correspondence without letting response times or service quality suffer. The challenge is not only detecting AI use; it is managing demand while preserving access to public services.

How Public Bodies Are Responding

The Information Commissioner’s Office published guidance in May after public authorities reported higher volumes and more complex FOI requests, some of which misquoted legislation or needed significant clarification. The ICO’s guidance, as described in the Sky News report, says an AI-written request remains a valid request. Authorities do not need to determine whether a person or an AI system wrote it, and they can seek clarification when a request is poorly formed.

Some organisations have tried a different approach. The Financial Ombudsman Service found that about a third of its complaints were AI-generated and began charging some professional representatives to bring complaints. The number of complaints fell, according to the source material. That example shows that procedural changes can reduce submissions, but it does not establish whether the submissions deterred were unnecessary or justified.

Sky News also cited a study by Chris Schmitz documenting 84 possible cases of “agentic flooding” across 11 jurisdictions. The report says many were roughly flat before 2022 and rose afterward. Examples included the UK Housing Ombudsman, where complaints increased from about 2,600 in 2022 to just over 7,000 in 2025, and the US Consumer Financial Protection Bureau, where complaints grew fivefold. These cases provide international context, but they do not establish a single cause for the increases.

“It’s like being bombarded with correspondence at speed.”

— Civil servants surveyed in research cited by Sky News

What the Rising Counts Cannot Prove

The reported timing is not proof of causation. The increases occurred after ChatGPT’s launch, but the source material does not establish how much of the change is due to generative AI, nor does it rule out other factors such as service problems, policy changes or shifts in public awareness. The figures are presented as growth since launch without detailed measurement periods and raw totals for every category.

The detector test also has clear limits: it covered 100 recent FOI requests, used a tool that can misclassify writing and does not establish who wrote each request. The survey finding that 58% of civil servants reported a significant impact is a measure of staff experience, not a count of AI-generated submissions. It is not clear how many additional submissions are AI-assisted, fully generated, or simply written with help from digital tools.

Nor is the balance between justified and unjustified claims settled across all services. The DWP figures and Schmitz’s research offer evidence against assuming that higher volume equals abuse, but they cannot show that every new complaint or claim is valid. The scale and administrative cost of the trend remain uncertain.

Redesigning Forms and Claims

The immediate response from public bodies is likely to involve applying existing rules to more correspondence and asking for clarification when requests are unclear. The ICO guidance indicates that FOI requests cannot be rejected simply because AI may have helped write them. How individual departments will manage workloads, maintain response times and assess complex submissions is not specified in the report.

Loosemore argues for changing the design of government services rather than making it harder for people to submit claims. His proposed direction is to use structured data instead of relying on essays wherever possible, so applicants can provide relevant information through clearer processes. He warned that friction-based restrictions may be bypassed as AI tools develop. The report does not set out an adopted government plan or a timetable for such redesign.

A further shift could come from AI agents that act on a person’s behalf, rather than merely drafting text for them to submit. Manthorpe demonstrated an agent service from Meta and raised examples such as appealing parking tickets or filing rail delay claims automatically. Whether such systems become widely used, and how authorities will handle the resulting submissions, remains open. The next test will be whether public services can absorb more requests while keeping legitimate access intact.

Key Questions

What has increased in Britain since ChatGPT launched?

Sky News reported increases across several written channels, including FOI requests, complaints to public bodies and employment tribunal claims. Reported growth varied by category, from 23% for complaints about public services to 144% for household complaints about water companies. The source material does not provide detailed measurement windows and raw totals for every figure.

Does the reporting prove AI caused the increase?

No. The report identifies a rise beginning around the time ChatGPT launched, but that timing does not prove causation. Manthorpe’s detector test was small and used technology that can be unreliable; other factors may also affect complaint and claim volumes.

Are AI-written FOI requests valid?

According to guidance from the Information Commissioner’s Office described by Sky News, AI-written requests remain valid FOI requests. Public authorities do not have to determine who or what wrote them, though they can ask for clarification if a request is unclear.

Are the additional complaints mostly frivolous?

The available evidence does not support treating all additional submissions as frivolous. Sky News reported that close to 40% of DWP complaints were upheld, down from the 50% range, and researcher Chris Schmitz said most cases he examined involved people claiming something they were entitled to. Those findings do not establish the validity of every complaint across all services.

What changes are being proposed?

Tom Loosemore argues that public services should rely less on long written accounts and use structured data and clearer processes instead. The report does not identify a formal government programme or timetable for making that change.

Source: ThorstenMeyerAI.com