AI Won't Save the State Time: Public Use of AI Is Flooding Institutions and Shifting Workload Elsewhere
Source: https://www.theguardian.com/profile/gabyhinsliff. "Don’t be fooled. The AI that was supposed to save us time is doing just the opposite | Gaby Hinsliff | The Guardian." October 20, 2026. www.theguardian.com
The Gist
AI makes it cheap and easy for everyone to fire off complaints, claims, letters and applications, so public bodies are being buried in machine-written material that humans still have to read. Like email and cars, a shortcut stops helping when everybody uses it. So promises that AI will save government lots of time and money should be doubted.
Conclusion
Government claims that AI will transform the state and save substantial time and public money should be treated with skepticism, because widespread AI use by the public generates floods of machine-generated input that cancel out or displace the efficiency gains.
Premises
- Since the adoption of ChatGPT, governments and official bodies worldwide appear to be experiencing unusual surges in complaints, requests and responses.
- AI-assisted submissions are visibly clogging public institutions: employment tribunal cases rose 39% in a year partly due to AI-drafted claims, courts face incoherent AI-built cases, and government departments and MPs' offices are swamped with AI-generated FOI requests and letters.
- These AI-generated inputs still require human review, because they may contain genuine grievances or come from vulnerable people entitled to a personal response, so using AI to sift them out is not a full solution.
- When both sides automate (as with AI-sifted job applications met by AI-mass-produced applications), neither side can separate signal from noise, producing paralysis rather than efficiency.
- Shortcuts stop being shortcuts once everyone uses them and can become bottlenecks, as with email overload and cars in cities, so wider adoption of agentic AI by the public is likely to worsen congestion.
- Time saved in one part of public services tends to reappear as extra work in another.
Assumptions
- The observed surges in complaints and requests are substantially caused by AI use rather than other factors such as economic pressures, legal aid cuts, or policy changes.
- Government savings projections do not adequately account for demand-side effects of public AI use.
- Human review of public submissions cannot be safely or ethically replaced by automated triage.
- Analogies from email and traffic congestion transfer to AI and public services.
- Early evidence ('early days') is indicative of the longer-term trend, and mitigating frictions will not be implemented quickly enough to prevent the costs.