AI Is Quietly Swamping Government Complaint Lines and Benefit Offices

Key points

• UK housing ombudsman complaints rose from 2,600 in 2022 to just over 7,000 last year, a near-tripling that tracks the spread of AI chatbots.
• US Consumer Financial Protection Bureau complaints grew fivefold over the same period.
• Researcher Chris Schmitz identified the pattern across 84 cases in 11 jurisdictions, calling it "agentic flooding."
• Most of the new filers appear to be legitimate claimants who previously gave up because the paperwork was too hard.
• Schmitz's paper is due at the AI Ethics and Society conference next month.

Something unusual started happening to government complaint offices and benefit services around 2022. The queues got much, much longer.

In the UK, the housing ombudsman, the body that handles disputes between tenants and landlords, went from roughly 2,600 complaints a year to just over 7,000. The Consumer Financial Protection Bureau in the United States, which fields grievances about banks and lenders, saw its complaint volume multiply by five. Brazilian courts recorded similar jumps in formal petitions, as did German parliamentary offices.

The timing matches the arrival of easy-to-use AI chatbots almost exactly.

What is actually going on?

Researcher Chris Schmitz calls the pattern "agentic flooding," a term for what happens when AI tools, software that can handle multi-step tasks like reading a letter and drafting a formal reply, make filing a claim far cheaper and easier.

His paper, first reported by TechCrunch AI and due for presentation at the AI Ethics and Society conference next month, covers 84 cases across 11 countries. In most of them, submission numbers sat roughly flat before 2022, then climbed steeply with no sign of stopping.

Schmitz stops short of declaring flat causation. But the curve is hard to ignore.

"People are finding out that this is something one can do," he said. "Before, it might have been a question of dragging a lot of context together and prompting ChatGPT 3.5 very precisely. Now it may be a question of just taking a photo of a letter with your Claude app and getting a pretty good response in one shot."

That single quote explains a lot. Filing a formal complaint used to demand time, confidence, and fluency in bureaucratic language. Many people with perfectly valid cases simply quit. Now they don't have to.

Should public services be worried?

The concern is real, but the situation differs from what hit corporate bug-bounty programmes, where companies invite outside researchers to find security flaws in their software. Those inboxes filled with low-quality, AI-generated junk reports last year. Staff still had to read every one.

Public services could face the same drain on staff time. Five times the applications on the same budget is a genuine problem, whatever the quality of the claims.

Schmitz's data points in a more hopeful direction, though. Unlike bug-bounty spam, the vast majority of new public-service filers appear to hold legitimate claims. They aren't bots gaming the system; they're people who were entitled to something and previously gave up because the process defeated them. Policy researchers call that barrier "administrative burden," the hidden cost in time and effort that stops eligible people from accessing services they already qualify for. We first picked up this concept on 10 September in our coverage of AI's effect on public services.

AI is now lifting that burden at scale.

What happens next?

Schmitz sees this as an opening, not just a headache. Governments that respond only by raising barriers to filing will lock out the same people AI just helped in. A more constructive path, he argues, is to redesign the services themselves for a world where assisted filing is the norm.

"Anyone who's ever used ChatGPT to do a tax return knows that there's a good version here where you're being helped," he said. "This could be the moment to say, we need to rethink pretty much everything about how this process looks."

That rethinking has barely started. Growth in filings shows no sign of slowing.

The thing a reporter on this beat keeps coming back to: the real story isn't the volume spike. It's that millions of people had valid claims and a system designed hard enough to stop them. AI broke that lock. Whether governments treat that as a crisis or a correction will tell us more about their priorities than any policy paper will.

Watch for: If you use an AI app to file a complaint or benefits claim, keep a copy of everything you submit. Agencies under pressure may take longer to respond, and a clear paper trail protects you if anything gets lost in the backlog.