Policy

UK Police Officer Used AI to Draft 'Tearful' Witness Statement

Court of Appeal rejects sentence appeal after officer instructed Microsoft Copilot to create statement designed to secure maximum penalty.

Omega Editorial· September 22, 2026· 3 min read

UK Court Confronts AI-Generated Evidence in Criminal Case

A UK police officer instructed Microsoft Copilot to draft a witness statement designed to "make a judge or reader weep" and secure the maximum sentence in a domestic violence case, according to court documents revealed this week. The officer has since been removed from front-line duties.

The Court of Appeal disclosed the incident while rejecting a sentence appeal by Jerome Gibson, 23, who was jailed for 44 months in 2024 after threatening his partner with a knife and causing grievous bodily harm to a police officer. Gibson's legal team argued the AI-generated statement improperly influenced the sentencing judge to assign greater weight to the harm caused.

How the Officer Used AI

According to testimony before the Court of Appeal, the officer made multiple entries into Microsoft Copilot while crafting the statement, which was presented as being written by Gibson's partner. The officer explicitly instructed the AI to "create the most impactful personal statement" and told it: "I want you to create it so that it meets the higher harm category."

The officer further directed the AI that its "goal is to reach the highest sentence," according to Mr Justice Lavender, who presided over the appeal alongside Lord Justice Edis and Mr Justice Linden.

Lucky Thandi, representing Gibson, told the court the officer had "made great efforts to ensure that the victim personal statement had read as powerfully as it possibly could."

Court Rules Sentence Stands

Despite the concerning conduct, the Court of Appeal determined the sentence was not "manifestly excessive." Justice Lavender noted that while the officer's use of AI was "deplorable," the facts presented were not "AI-generated hallucinations" but rather statements the victim had made during her police interview.

"It does not appear that the use of AI, deplorable though it is, caused the appellant's sentence to be any longer than it otherwise would have," Lavender said.

Denis Barry, representing the Crown Prosecution Service, acknowledged the officer's conduct was "concerning" but argued that an experienced judge would know how to evaluate multiple statements appropriately. Barry emphasized that "facts in criminal cases speak for themselves and they do not require exaggerated language in criminal proceedings."

Why It Matters

This case exposes a critical gap in law enforcement protocols around generative AI use. While the court found no material impact on sentencing in this instance, the officer's explicit instructions to manipulate emotional impact raise serious questions about evidence integrity and due process. As AI tools become ubiquitous in professional workflows, criminal justice systems worldwide will need clear guidelines distinguishing legitimate administrative assistance from improper manipulation of evidence presentation. The incident also highlights how AI can amplify existing human biases—in this case, an officer's desire to secure maximum punishment—rather than simply automating neutral tasks.

The details of the case were first reported by the BBC, which covered the Court of Appeal proceedings in Nottingham.

#artificial intelligence#law enforcement#microsoft copilot#criminal justice#evidence integrity#uk courts

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Policy

Policy· 3 min read

Trump to Name AI Czar as US-China Competition Intensifies

The president announced plans for an AI envoy and task force while resisting calls to slow development, ahead of a pivotal meeting with Xi Jinping.

Via AI Watch · Sep 22, 2026
Policy· 2 min read

UN Secretary General Demands Global AI Safety Standards

António Guterres warns against autonomous weapons and calls for accountability frameworks as world leaders gather in New York.

Via AI Watch · Sep 22, 2026
Policy· 3 min read

$68 Billion in U.S. Data Center Projects Stalled by Local Opposition

Community resistance blocked or delayed 45 facilities between April and June as concerns mount over AI infrastructure's environmental and economic impact.

Via AI Watch · Sep 22, 2026