Could Friction Slow The AI Paperwork Flood In Britain?

🔍 Read the full analysis: Could Friction Slow The AI Paperwork Flood In Britain? on ThorstenMeyerAI.com

Written complaints and claims to UK public bodies have increased across multiple services since ChatGPT launched in November 2022, according to a Sky News report published October 20. The report links the pattern to the lower effort of producing formal submissions with AI, while stressing that the figures do not establish AI as the cause—and that many people using the systems may have valid complaints or claims.

Data reviewed by Sky News showed increases in a range of channels, including household complaints about water companies, up 144%, complaints to English local authorities, up 78%, and written complaints to GP practices, up 71%. Complaints to the Department for Work and Pensions rose 64%, while individual employment tribunal claims rose 62%. The report gives these as growth since ChatGPT launched; the supplied figures do not specify a common measurement window or provide a more detailed baseline for each channel.

Other increases in the report included complaints about legal services, up 49%; written complaints to the NHS, up 35%; complaints to police forces, up 34%; and complaints about public services, up 23%. Sky News described a similar pattern across the channels: activity was broadly flat for years, then began rising after late 2022. Each service has its own possible drivers, and the report notes that water companies, for example, have faced a difficult period.

Sky technology correspondent Rowland Manthorpe also tested 100 recent Freedom of Information requests with an AI-text checker. The tool classified 89% as fully AI-written and 9% as human; a smaller sample from 2019 was classified as entirely human. Manthorpe cautioned that AI detectors are unreliable and that this small test cannot prove what caused the wider increase.

Pressure on Public-Service Inboxes

The increase matters because many UK services rely on written forms, letters and appeals to decide what help or remedy someone receives. If AI makes these submissions easier to prepare, staff may face more correspondence and more complex cases, while the systems handling it may not be designed for a sharp rise in volume. A survey of civil servants cited by Sky News found 58% said AI-generated material from the public significantly affects their work.

But treating volume as evidence of abuse risks turning away people with legitimate concerns. Sky News reported that the share of DWP complaints upheld has fallen from the 50% range to close to 40%. That is a lower proportion, not an absence of valid cases; the report says the larger overall volume still means more upheld complaints in absolute terms. The central policy choice is whether to discourage submissions or improve the process for sorting and resolving them.

The challenge may grow as AI tools move from drafting text to taking actions for users. Automated systems could submit appeals or claims at scale, including for parking tickets or delayed rail journeys. If those claims are eligible, the issue is not simply unwanted correspondence: it is whether public systems can process valid requests fairly and promptly.

From Chatbots to Automated Claims

In May, the Information Commissioner’s Office issued guidance after public authorities reported higher volumes of AI-generated Freedom of Information requests, including some that misquoted legislation or needed clarification. The ICO said AI-written requests remain valid, authorities do not need to identify whether AI produced them, and they can ask for clarification when a request is unclear.

The trend is not confined to Britain. Researcher Chris Schmitz documented 84 possible cases of “agentic flooding” across 11 jurisdictions, in a paper reported by TechCrunch in September. The source material says many cases were broadly flat before 2022 and increased afterwards. It also cites the UK Housing Ombudsman, where complaints rose from about 2,600 in 2022 to just over 7,000 last year, and the US Consumer Financial Protection Bureau, where complaints grew fivefold.

Some organisations have tried adding barriers. The Financial Ombudsman Service found about a third of its complaints were AI-generated and began charging some professional representatives to bring complaints; the number of complaints subsequently fell, according to the report. That example shows friction can reduce volume, but by itself does not establish whether it filters invalid claims or deters people with valid ones.

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

— Civil servants surveyed in research cited by Sky News

Cause and Claim Validity

The reported timing and broad pattern do not prove that ChatGPT caused the increases. The figures cover different services with different circumstances, and the source material does not set out a shared measurement period or explain every channel’s baseline in detail. Other factors, including service-specific problems and changes in public awareness, may also contribute.

The AI-text checker test is limited: detection tools can misclassify writing, and a sample of 100 recent requests cannot establish how much of the overall rise was caused by AI. Nor do the figures show how many submissions were produced with assistance, how many were fully automated, or how many were upheld across all services. The report presents the DWP and Financial Ombudsman examples, but they cannot settle the picture for every public body.

Redesigning Forms and Appeals

The immediate response from public bodies is likely to focus on handling more requests while applying existing rules. The ICO’s guidance says authorities can seek clarification but cannot reject a Freedom of Information request merely because AI may have helped write it. The source material does not identify a new national policy or a timetable for changes to complaint and appeal systems.

Loosemore’s proposed alternative is to move away from essay-style applications and towards structured data and clearer digital processes. He warned that adding barriers could trigger a contest with technology providers able to work around them. The report leaves open how such services would be funded, designed and introduced, and whether they could reduce staff workload without making access harder for people entitled to assistance.

For now, the key test is whether public bodies can separate valid requests from unclear or ineligible ones without treating AI use itself as misconduct. The scale of any shift to agents that submit claims automatically, and its effect on response times and outcomes, remains to be seen.

Key Questions

What did the Sky News report find?

It found rising numbers of written complaints, Freedom of Information requests and employment tribunal claims across several UK services since ChatGPT launched in November 2022. Reported growth varied by channel, from 23% for complaints about public services to 144% for household complaints about water companies.

Does the report prove AI caused the rise?

No. The timing and similar patterns are consistent with AI making formal submissions easier, but the report does not establish causation. Different services may also have experienced other changes that affected complaint volumes.

Are AI-written Freedom of Information requests valid?

The Information Commissioner’s Office said in May that AI-written requests are still valid requests. Authorities do not need to determine who or what wrote them, though they can ask for clarification if a request is poorly formed.

Are the additional complaints mostly bogus?

The report does not establish that. It cites researcher Chris Schmitz’s view that most cases in his research involve people claiming something they are entitled to claim. The figures do not provide a full validity assessment across all the UK services discussed.

What changes are being discussed?

The report contrasts adding friction, such as charges or more demanding processes, with redesigning services around structured information rather than long written submissions. It does not identify a confirmed national plan or implementation timetable.

Source: ThorstenMeyerAI.com