UK public bodies face a rising flood of chatbot-written complaints
UK MPs, councils, lawyers and schools report a surge in AI-generated complaints and legal challenges that are slower and harder to handle.
According to The Guardian, public bodies across the United Kingdom are struggling with a sharp rise in correspondence written with artificial intelligence tools. The paper uses the phrase "AI clog" to describe the steady stream of grievances, information requests and legal challenges reaching Members of Parliament, local authorities, lawyers and schools.
The report says the problem was small in 2024 and has since grown into what one worker called an onslaught. Hollie Wickens, a senior caseworker in a Labour MP's office, told The Guardian that about half of the emails her office receives now involve some form of AI.
Staff say this material is often complex to read and difficult to answer, which adds time to tasks that were already demanding. Wickens said the work has become more taxing, even though she has worked in the role for a decade.
The Guardian's account rests mainly on first-hand descriptions from people who handle this correspondence every day. The report does not present formal national counts, so its picture of the trend comes from those working in the front line of public offices.
Background
Public bodies in the UK have long received large volumes of written correspondence, from individual letters to organized campaigns. Freedom of information laws give members of the public the right to request records from authorities, and those requests must be handled within set legal timeframes. Local councils, schools and government departments also face formal complaints procedures, and many of these processes require a written response that is checked for accuracy and legal risk.
Mass correspondence is not new. Campaign groups have used template letters and email tools for years, and offices have developed ways to group similar messages. What has changed is the cost of producing text. Generative AI tools, widely available since late 2022, can draft long, polished and personalized letters in seconds. A complaint that once took real effort to write can now be produced quickly and copied many times.
This shift affects the balance of effort between the sender and the recipient. Writing a message takes moments, but reading, checking and responding to a long or legally framed one can take far longer. Staff may need to determine which claims are genuine, which are repeated, and which raise legal issues that need specialist review. Smaller offices with limited staff can feel this imbalance most sharply.
Why it matters
The rise of AI-assisted correspondence raises practical questions about fairness and access. If public offices are swamped, cases from residents who write in their own words could face slower replies. A system built on the assumption that a written complaint reflects a person's direct experience may be strained when text is generated at scale.
There is also a risk of cynicism. If staff begin to assume that any long or polished message is automated, genuine concerns may be treated with less care. Officials must avoid that trap while still managing workloads that may be growing faster than resources. The problem is not simply that people are using new tools. It is whether public institutions can keep their promises of timely, reasoned responses.
The trend also matters for legal processes. Lawyers and courts depend on accurate filings, and an influx of challenges that are hard to assess could add to delays. How public bodies triage, verify and answer this material may shape how well they serve the people who depend on them.
What to watch
Watch whether public bodies publish clearer guidance on handling AI-assisted correspondence, including how they check accuracy and assess urgency. Also watch whether more detailed data emerges. The current picture relies on first-hand accounts, and firmer figures would show how widespread the problem is across different sectors.
Pay attention to how courts and tribunals respond to legal challenges that appear to be drafted with AI tools. Finally, consider whether offices adopt new sorting or summarizing tools, and whether those tools themselves raise questions about transparency and error.
Viewpoints
Commentary from our columns: three angles on this story.
The User what changes for people who use it
If you have ever written to your MP, a council or your child's school, you may have wondered whether anyone actually reads your message. The Guardian's reporting suggests that the people receiving these letters are now sorting through a growing pile of text, and that some of it is hard to tell apart from routine correspondence. For ordinary residents, the risk is simple: a sincere complaint about a bin collection, a benefit decision or a classroom issue could sit in a queue behind a flood of generated messages. The people who suffer most are often those with the least time to chase a reply. Public services should make it easy to ask for help without requiring a polished letter or a lawyer. Offices may need clearer ways to show that a message has been received, reviewed and assigned. Residents, meanwhile, can keep it direct: state the problem, give the facts they know, and say what outcome they want. A short, honest message from a real person still deserves a real answer, and institutions should be judged on whether they still deliver one.
The Skeptic open questions and weak points
The headline phrase is vivid, but the evidence so far is mostly a set of first-hand impressions. One caseworker's estimate that half of her office's emails show some AI involvement is striking, yet it is hard to know how that was measured, how confident anyone is in spotting machine-written text, or whether the pattern holds elsewhere. Writing that is long, formal or well organized is not proof of automation; people have always used templates, advocacy groups and professional help. There is also a risk in treating any polished complaint as suspect, since that could disadvantage the very people who most need help writing clearly. Before concluding that AI is clogging public institutions, we should ask for counts, comparisons with earlier years and an account of what changed in workloads beyond the volume of messages. Strong claims deserve strong evidence, and the current reporting, while useful, does not yet supply it.
The Rulebook privacy, safety and regulation
Public bodies receiving large volumes of correspondence are bound by duties that do not disappear because the messages are generated. Authorities must still handle requests under the rules that apply to them, and they must still weigh personal data and legal claims with care. The rise of AI-written submissions raises questions about how those duties are met. Can staff verify identities and claims when text can be produced at scale? How should a body respond to a challenge that may contain inaccurate or invented citations? There is also the question of what offices themselves do with incoming material. If they feed correspondence into automated sorting or summarizing tools, they need clear rules on data handling, accountability for errors and human review of decisions. Good regulation here would not ban the use of AI by the public. It would set expectations for how institutions process what they receive, make clear who answers for mistakes, and protect people whose information passes through these systems.