A UK officer is investigated for using AI to 'create evidence.' The mundane version is scarier
Derbyshire Police are investigating one of their own for allegedly using AI to create evidence across multiple cases. The force won't say what the material was — and the thread's most plausible read isn't deepfakes.
The phrase conjures a fabricated video of a suspect. A more likely scenario, one commenter argues: an officer 'enhanced' a blurry photo, letting AI invent the missing detail and convincing themselves it was harmless.
The deeper worry runs past one officer. Once anyone can generate images and video on demand, readers ask whether whole categories of police evidence quietly become unreliable — and how many convictions already lean on them.
What's actually alleged, and what isn't
Sky News reports that Derbyshire Police are investigating an officer for allegedly using AI to "create evidence" in multiple cases, and that the force declined to detail what the evidential material consisted of. That vagueness is where the thread did its work. A reader who surfaced a fuller writeup noted how loose the key term is:
"[The Derbyshire Police] declined to give more detail about what the evidential material consisted of. The term [evidential material] can be used to describe witness statements." — a reader citing a fuller FT account
The boring scenario is the dangerous one
"Create evidence" sounds like a fabricated crime video. The thread's most developed reading is far more ordinary, and for that reason more worrying — because it's the kind of thing a person can talk themselves into:
"what I think is much more likely is the police officer used AI to enhance an image in a way that they considered innocuous… a photo was blurry so they 'enhanced' it. Since 'enhancing' is letting AI fill in the gaps it would be using AI to 'create evidence'." — a commenter walking through the innocuous-looking version
That distinction matters in court. A forged video is a discrete crime. An "enhanced" photo, where a model hallucinates the detail that makes an ID stick, can pass as routine processing while doing the same damage.
The systemic question the case opens
Several readers zoomed out from one officer to the whole evidentiary regime in an age of generative media:
"i do wonder, that in the age where we have image and video creation out of the bag, whether or not this will result in whole classes of evidence becoming completely unreliable." — a reader on evidence categories going dark
"I wonder how many people have been unjustly imprisoned between planted evidence, made up evidence, and illegal parallel construction…" — a reader connecting it to older evidence-integrity failures
Chain of custody meets the fill tool
The lasting problem here isn't one investigator. It's that "enhancement" tools now generate plausible detail that was never captured, and the line between cleaning up an image and inventing one has blurred to nothing. Courts have spent a century building chain-of-custody rules for physical evidence. This case is an early sign those rules have no answer yet for material that looks photographic but is partly synthesized — and that the most corrosive misuse won't announce itself as a deepfake. It arrives wearing the label "enhanced," looking like routine work.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
A guide to 'AI coding without going broke' gets schooled by people who go broke far slower
The post pitches three ways to code with AI at home and claims $400 a month of plans buys $2,800 of API usage. Commenters who do this daily report spending cents, and dismiss most of the advice as 'pay for max.'