Britain’s Paperwork Pile Grows As AI Spreads

🔍 Read the full analysis: Britain’s Paperwork Pile Grows As AI Spreads on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

A Sky News report published October 20 found that written complaints, Freedom of Information requests and individual employment tribunal claims have risen across Britain since ChatGPT launched in November 2022, raising questions about how public services handle more submissions. The data show increases across multiple channels, but do not establish that AI caused them; a survey cited in the report found 58% of civil servants said AI-generated material from the public significantly affects their work.

Sky News technology correspondent Rowland Manthorpe compared activity across several types of written submissions. Since ChatGPT’s launch, the report found household complaints about water companies rose 144%, while complaints to English local authorities rose 78%. Written complaints to GP practices increased 71%, complaints to the Department for Work and Pensions (DWP) rose 64%, and individual employment tribunal claims increased 62%.

The report also recorded increases in complaints about legal services, the NHS, police forces and public services. The figures describe growth since the chatbot’s launch; they do not, on their own, identify how much of each increase came from AI use. Each channel has other possible causes. Sky News noted, for example, that water companies have faced a difficult period. The report’s argument rests partly on a similar upward pattern across otherwise separate services, rather than on proof of a single cause.

Manthorpe tested 100 recent FOI requests with an AI-text detector. It classified 89% as fully AI-written and 9% as human-written; a smaller 2019 sample was classified as entirely human. The report cautioned that detectors are unreliable and that this small test does not prove AI caused the broader rise. Separately, an academic survey cited by Sky News found 58% of civil servants said AI-generated public submissions have a significant effect on their work.

More Claims Test Public-Service Capacity

More written submissions can put pressure on teams that must read, clarify and answer them. The civil-servant survey responses cited by Sky News describe correspondence arriving quickly and at greater volume, with some workers saying it feels like responding to a chatbot. That reported strain matters because delays or rushed handling can affect people seeking answers, benefits, redress or an appeal.

But a rise in volume is not the same as a rise in baseless claims. Sky News reported that the share of DWP complaints upheld has fallen from the 50% range to close to 40%. Even at a lower success rate, a much larger total number of complaints can mean more upheld complaints in absolute terms. Researcher Chris Schmitz, whose work on possible cases of “agentic flooding” was reported by TechCrunch in September, said most cases he found involved people claiming something they were entitled to receive.

This makes the policy choice consequential: adding barriers may reduce workloads, but could also deter people with valid complaints. The challenge is to process submissions efficiently without treating AI use itself as evidence that a claim is false or undeserving.

Rules and Automation Are Already Shifting

In May, the Information Commissioner’s Office (ICO) issued guidance on AI-generated FOI requests after public authorities reported higher volumes and more complicated submissions. According to the guidance as described by Sky News, AI-written requests remain valid. Authorities do not have to identify whether a person or an AI system wrote them, though they can seek clarification when a request is unclear or badly formed. The report also says some AI-generated requests misquote legislation.

The issue is not limited to Britain. Schmitz’s paper, as reported by TechCrunch in September, identified 84 possible cases across 11 jurisdictions of what he calls “agentic flooding.” The UK Housing Ombudsman received about 2,600 complaints in 2022 and just over 7,000 last year, according to the report. The US Consumer Financial Protection Bureau saw complaints grow fivefold. These examples indicate a wider concern, but do not establish that all increases have the same cause.

At present, people often use chatbots to draft text and then submit it themselves. Manthorpe’s report also demonstrated an AI agent service from Meta, which can act on a user’s behalf rather than simply produce a draft. Automated systems could make it easier to pursue routine claims, such as parking appeals or rail delay compensation, raising new questions about how existing processes will cope.

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

— Civil servants surveyed in research cited by Sky News

The Cause of Growth Is Unproven

The reported increases do not show how many additional submissions were written with AI, nor what share of overall growth can be attributed to chatbots. The comparison period begins with ChatGPT’s November 2022 launch, but timing alone cannot establish causation. Other changes in public services, disputes and awareness of complaint routes may also affect volumes.

The FOI detector test is limited: AI-text checkers can misclassify writing, and the recent sample of 100 requests cannot represent every request. The figures provided do not include enough detail to assess all underlying comparisons, such as the full count for each channel or the precise periods used. It is also unclear how many submissions were generated wholly by AI, edited with AI assistance, or written without it.

Nor do the figures settle whether increased volume is mostly administrative burden, newly accessible legitimate demand, or a combination. The DWP upheld-complaint rate and Schmitz’s assessment offer relevant evidence, but do not describe every agency or type of claim.

Service Design Faces the Next Test

The next test for public bodies is how they handle higher volumes without excluding people who have a valid reason to complain or appeal. The ICO’s existing guidance says AI-written FOI requests remain valid and permits authorities to seek clarification where needed. The report does not identify a new national rule or government response announced alongside its findings.

Loosemore argues for changing the design of public services rather than adding obstacles to submissions. He favours systems based on structured data instead of processes that require people to write essays. He warned that attempts to add friction could be bypassed as AI tools develop, and said he could not see the current way UK benefits are run surviving without fundamental changes. Those are his assessments, not a government decision.

As AI agents become more capable of submitting or pursuing claims for users, agencies may face additional demand for routine processes. What remains to be seen is whether public services will adapt their forms and decision systems, increase capacity, or introduce other controls. The report leaves that policy response open.

Key Questions

What did the Sky News report find?

It reported rising written submissions across several UK channels since ChatGPT launched in November 2022, including a 144% increase in household water-company complaints and a 64% increase in DWP complaints. The figures do not prove AI caused those increases.

Did the report prove that AI caused the rise?

No. The report found a similar upward pattern across multiple channels, but said the AI-detector test was not conclusive. Other factors may have contributed to the growth.

Are AI-written FOI requests valid?

According to ICO guidance published in May, AI-written requests are still valid FOI requests. Public authorities do not need to determine who or what wrote them, and can ask for clarification if a request is unclear.

Are the additional complaints mostly baseless?

The available evidence does not support that conclusion. Sky News reported that the DWP’s upheld share has fallen to close to 40%, while researcher Chris Schmitz said most cases in his research involved people claiming something they were entitled to. Neither finding establishes the merits of every submission.

What changes could public services make?

Tom Loosemore has argued for services based on structured data rather than long written applications. The report does not say the government has adopted that proposal or announced a new nationwide response.

Source: ThorstenMeyerAI.com