AI Is Flooding Britain’s Bureaucracy With Documents, And Friction Won’t Stop It

🔍 Read the full analysis: AI Is Flooding Britain’s Bureaucracy With Documents, And Friction Won’t Stop It on ThorstenMeyerAI.com

A Sky News report published October 20 found that written complaints, Freedom of Information requests and employment tribunal claims have risen across several British public services since ChatGPT launched in November 2022. The increase is putting pressure on systems that rely on people writing applications and appeals, though the report does not establish that AI caused the rise or that the additional submissions are mostly invalid.

Technology correspondent Rowland Manthorpe examined data across complaint channels involving water companies, local authorities, GP practices, the Department for Work and Pensions, legal services, the NHS and police forces. The report said the pattern in many channels was a long period of little change followed by increases beginning around late 2022. It gave growth since ChatGPT’s launch as 144% for household complaints about water companies and 78% for complaints about English local authorities. Other reported increases included 71% for written complaints to GP practices, 64% for complaints to the DWP and 62% for individual employment tribunal claims.

Those figures are changes in the channels Sky News tracked; they do not by themselves show how much of the increase was caused by AI. The report noted that each service has its own pressures. Water companies, for example, have faced difficult years. In a small test, Manthorpe ran 100 recent Freedom of Information requests through an AI-text detector. The tool labelled 89% fully AI-written and 9% human. A smaller 2019 sample was labelled entirely human, but the report cautioned that AI detectors are unreliable and the test cannot prove why request volumes rose.

A survey of civil servants cited by Sky News found that 58% said AI-generated material from the public significantly affects their work. Separately, the Information Commissioner’s Office issued guidance in May after authorities reported higher volumes and more complex FOI requests. The ICO said AI-written requests remain valid, authorities do not need to establish who or what drafted them, and officials can seek clarification when requests are unclear.

Pressure on Paper-Based Services

The increase matters because many public services still require people to explain a problem in writing, then rely on staff to read and assess those submissions. Generative AI can lower the effort needed to draft a complaint or appeal, while AI agents may eventually submit forms or claims on a person’s behalf. That could increase workloads and make existing delays harder to manage.

But volume alone is not evidence of abuse. The report cited researcher Chris Schmitz, whose analysis of possible cases of what he calls “agentic flooding” found that most cases involved people claiming something they were entitled to claim. At the DWP, the proportion of complaints upheld reportedly declined from the 50% range to close to 40%. Because the total number of complaints grew, Sky News argued that the absolute number upheld could still be higher. The figures point to a service-design challenge: reducing submissions may also block people with valid complaints.

From Chatbots to Automated Claims

Sky News described a similar broad pattern across several separate channels, but the evidence presented is not a controlled test of AI’s effect. The timing follows ChatGPT’s public launch, yet other causes may contribute to each channel’s trend. The report also referred to a paper by Chris Schmitz, covered by TechCrunch in September, which documented 84 possible cases across 11 jurisdictions. It reported, among other examples, that complaints to the UK Housing Ombudsman rose from about 2,600 in 2022 to just over 7,000 the following year, while complaints to the US Consumer Financial Protection Bureau grew fivefold.

For now, much AI-assisted correspondence is drafted by people using chatbots. The next change may come from AI agents that can take actions, not just produce text. Manthorpe demonstrated an agent service from Meta in the report. If such tools can submit routine claims automatically, public bodies may face more applications arriving with little additional effort from each applicant.

The Financial Ombudsman Service, according to Sky News, found that about a third of its complaints were AI-generated and began charging some professional representatives to bring complaints; complaint numbers then fell. That example shows one possible response, but it does not establish that charges are appropriate across public services or that reduced volumes mean fewer valid cases.

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

— Civil servants surveyed in research cited by Sky News

Cause and Claim Validity

The available figures do not isolate AI as the cause of rising complaints or requests. The AI-detector test was small, and the report itself warned that detection tools can be unreliable. The source material does not provide the underlying datasets, full survey methodology or a breakdown showing what share of all new submissions used AI.

It is also unclear how many additional submissions are valid across each service, how much extra staff time they require, and whether AI agents will substantially change volumes. The DWP figures suggest the share of complaints upheld has fallen, but do not establish the overall validity of claims across other agencies. Trends may differ by service and may reflect changes unrelated to AI.

Redesigning Applications and Appeals

The immediate policy choice is whether public bodies add friction—such as fees or barriers to AI-drafted submissions—or redesign processes that currently ask people to write essays. The ICO’s existing guidance says authorities should treat AI-written FOI requests as valid and seek clarification where needed, rather than trying to identify the authoring tool.

Loosemore has argued for services built around structured data rather than long written accounts. The source material does not set out a government decision, timetable or specific reform plan. Public bodies will need to assess their own workloads and safeguards as AI agents develop, while ensuring that changes intended to manage volume do not prevent people from pursuing legitimate claims.

Key Questions

What did the Sky News report find?

It reported rising written complaints, FOI requests and claims across several British public services since ChatGPT launched in November 2022. The increases vary by channel, and the report does not prove AI caused them.

Does the evidence show that most new complaints are AI-generated?

No. A small detector test labelled most of 100 recent FOI requests as fully AI-written, but the report cautioned that AI detectors are unreliable. It did not establish what share of all new complaints or claims were produced with AI.

Are AI-written FOI requests valid?

The Information Commissioner’s Office said in guidance published in May that AI-written requests remain valid. Authorities do not have to determine who or what wrote them, and can ask for clarification if a request is unclear.

Why might higher claim volumes reflect unmet need?

AI can make it easier for people to write formal complaints or appeals. The report cited researcher Chris Schmitz’s view that most cases in his analysis involved people claiming something to which they were entitled, although that finding does not establish the validity of every submission.

What response is being proposed?

Tom Loosemore argues that public services should move away from essay-based applications toward structured information. The source material reports no government decision or timetable for such a change.

Source: ThorstenMeyerAI.com