An investigative report published by The Guardian revealed that public bodies, local authorities, and parliamentary offices in the United Kingdom are suffering from an “AI clog” phenomenon, resulting from a flood of thousands of lengthy complaints and requests drafted by citizens via generative chatbots, which sometimes reach up to 200 pages and include inaccurate legal citations that drain reviewers’ time.
The administrative clog phenomenon and the growth of automated correspondence in Britain
The imbalance between the ease of text generation and the cost of institutional review
The inflation of complaint sizes to 200 pages and legal hallucinations in texts
Official statistics at Transport for London and employment tribunals
The trade-off between automated sorting and protecting citizens’ right to file complaints
Frequently asked questions
The administrative clog phenomenon and the growth of automated correspondence in Britain
An investigative report published by The Guardian on November 8 at 03:00 AM EST, corresponding to 12:00 PM Gulf Standard Time, revealed that public institutions, employers, and professional organizations in the United Kingdom are struggling to handle an ever-increasing volume of long and complex correspondence prepared using generative AI tools.
This worsening administrative phenomenon is known as the “AI clog.” The investigative report was based on original interviews with local authority lawyers, staff in the British Parliament, employers, and experts, along with supportive public records. The investigation focuses on the direct implications of AI on daily administrative workload, rather than the launch of a new technical model.
The imbalance between the ease of text generation and the cost of institutional review
Smart generative assistants allow members of the public to write official letters, administrative complaints, appeals, and information requests without specialized training or prior legal expertise, which could theoretically improve access to services for people who previously struggled to articulate their concerns and demands formally.
However, generating a long document via chatbots takes far less time than it requires for a competent employee to carefully review it to verify the facts, legal relevance, and determine the required corrective action. This sharp imbalance in time effort doubles the workload on public institutions legally obligated to provide thoughtful and accurate individual responses to every incoming correspondence.
The inflation of complaint sizes to 200 pages and legal hallucinations in texts
The chief executive of Lawyers in Local Government told The Guardian that correspondence that previously did not exceed about two pages is now about 10 times longer, with some individual complaints extending to nearly 200 full pages; noting that these figures represent field observations provided by a professional representative rather than audited national averages. Similarly, a casework officer in an MP’s office estimated that about half of the emails she receives contain AI-generated material, though this local estimate should not be automatically generalized to every parliamentary office.
The report also tracks the spread of legal fallacies and errors; language models can cite cases and precedents from entirely incorrect jurisdictions or fabricate details that appear superficially convincing but are inaccurate. Long documents with an assertive tone often overlook the core issue or fail to clarify what the sender actually wants, meaning that a formally polished complaint still requires fact-checking and a clear definition of the desired outcome, because the ease of text generation does not make legal reasoning reliable.
Official statistics at Transport for London and employment tribunals
Supporting administrative statistics show a tangible increase in operational pressures, even if direct causality is less certain; Transport for London received about 4,769 Freedom of Information requests during the 2025–2026 period, an unprecedented record, and the authority’s audit committee noted an increase in appeals that appeared to be drafted with AI assistance.
Employment tribunal claims also saw a notable rise during a previous period, though these changes cannot be entirely and automatically attributed to generative AI alone, given that legislative amendments, living economic conditions, and other social factors all influence the volume of demand for institutional and judicial services.
The trade-off between automated sorting and protecting citizens’ right to file complaints
Some organizations are currently exploring regulatory solutions including imposing stricter submission forms, setting shorter word limits for requests, or using AI-backed sorting systems. However, these approaches create sensitive trade-offs; overly strict automated filters may mistakenly reject a legitimate request, automated responses may worsen misunderstandings, and the UK Information Commissioner’s Office has emphasized that requests still require proper human processing even when the auto-generated text is repetitive or inaccurate.
The significance of this issue extends far beyond office productivity; generative tools shift the burden of effort from the writer to the recipient, especially when an individual can produce unlimited material while the institution is obligated to review it. The most useful future evidence will consist of well-designed studies measuring processing times, error rates, user accessibility, and the success of redesigned models in reducing unnecessary burden without excluding legitimate complaints.
Frequently asked questions
Question: What is meant by the AI clog phenomenon revealed by The Guardian?
Answer: It is the phenomenon of overwhelming British public institutions, municipalities, and parliamentary offices with a flood of long and complex complaints and correspondence easily generated by citizens via generative AI tools, draining staff time in reviewing them.
Question: How did the length of complaints submitted to local authorities change according to the report?
Answer: The chief executive of Lawyers in Local Government noted that correspondence that previously amounted to about two pages is now about 10 times longer, with some complaints reaching nearly 200 pages.
Question: What are the most prominent legal problems in complaints written by chatbots?
Answer: Language models frequently cite cases from wrong jurisdictions or fabricate convincing-sounding details, and lengthy documents may overlook the core issue or fail to specify the complainant’s actual demand.
Question: How many Freedom of Information requests did Transport for London receive?
Answer: Transport for London recorded a record 4,769 information requests during 2025–2026, with the audit committee noting a noticeable increase in AI-drafted appeals.
