UK court condemns police use of AI to generate victim impact statements
The Court of Appeal issued a stark warning after a police officer used Microsoft Copilot to create an emotionally charged statement intended to maximize a criminal sentence.
The short version
- A police officer used Microsoft Copilot to turn a victim interview into a 'tear-jerking' statement.
- The Court of Appeal described the use of AI to generate such statements as 'deplorable'.
- The court ruled that no police officer should ever prepare witness statements in this manner.
- The officer involved was removed from frontline duties and is currently under investigation.
- The ruling emphasizes that witness statements must be factual accounts, not rhetorical advocacy.
Why it matters
This sets a clear precedent that using AI to 'spin' evidence is a violation of judicial integrity. It highlights the risk of AI-generated content being used to manipulate legal outcomes rather than simply assisting with administrative tasks.
The 'tear-jerking' prompt
The Court of Appeal recently addressed a case where a police officer used Microsoft Copilot to draft a victim personal statement (VPS). The goal of the officer was to create a document that would maximize the emotional impact on the court. Specifically, the officer wanted the statement to influence the sentencing of an offender by focusing on the most severe aspects of the crimes.
According to the source, the officer uploaded a transcript of a victim's interview into the AI tool. They provided specific instructions to the AI to create the most impactful statement possible. The officer even told the AI to aim for a "higher harm category" for sentencing purposes. They explicitly requested a statement designed to make a judge or reader weep.
The resulting statement was criticized by the court for being an exercise in rhetoric rather than a factual account. While it did include some true details from the victim, the court noted it felt like advocacy. The judge in the original sentencing hearing recognized this and chose to rely only on the limited factual points the victim had actually recounted.
A 'deplorable' misuse of AI
The Court of Appeal issued a stern warning regarding the use of artificial intelligence in criminal proceedings. The court described the officer's actions as "deplorable." It emphasized that the issue was not the technology itself, but the way it was used to corrupt the evidential process. The court stated that no police officer should ever prepare a witness statement in this way.
The court noted that while there is no rule requiring a witness to type their own statement, there are strict requirements for how they are drafted. Statements must accurately reflect the witness's own uninfluenced evidence and be limited to their factual account. They should not contain unauthorized opinions or advocacy-style legal submissions that argue a specific point of law.
The ruling clarifies that the primary duty of a witness statement is to provide a truthful account of facts within the witness's own knowledge. It is not intended to provide commentary on other documents or engage in arguments. The court's response serves as a warning that using AI to bypass these standards is wholly unacceptable in any judicial forum.
Impact on the legal process
This case highlights a growing concern about the quality of AI-generated submissions in various legal settings. Some courts have already seen applications that are extremely lengthy, unfocused, and verbose. These submissions often repeat the same propositions in slightly different language without focusing on the actual issues before the court.
AI-generated content can also lead to the inclusion of 'hallucinated' or non-existent legal authorities. In some instances, the AI produces citations that are technically correct but entirely irrelevant to the point in issue. This creates a burden for judges who must sift through irrelevant information to find the core facts of a case.
The officer involved in this specific case was removed from frontline duties following the discovery of his conduct. He is currently under investigation regarding his actions in this case and others. This underscores that the misuse of AI in legal drafting can have significant professional consequences for those who facilitate it.
New rules and directions
Some coroners have already begun to implement specific directions to regulate the use of AI in their courts. These rules often require individuals to independently verify the accuracy of all factual submissions and legal authorities. They may also require a statement confirming that the author has personally reviewed the material to ensure it is not a product of AI hallucinations.
One proposed change involves amending the statement of truth to require a signatory to confirm the statement is their own. This would require individuals to declare whether they used AI to assist in drafting the document. Such a move would increase transparency regarding the origin of the evidence presented to the court.
The Court of Appeal's remarks are expected to apply with equal force across different jurisdictions. While AI can be a useful time-saving tool for research, the judiciary is signaling that it cannot be used to substitute for genuine witness evidence. The focus remains on maintaining the integrity of the judicial process against automated advocacy.
Maintaining evidence integrity
The court emphasized that the core requirement of a witness statement is that the witness must adopt the statement as true. This usually involves the witness reading the statement before signing it to confirm its accuracy. When AI is used to generate the text, there is a risk that the witness may not fully grasp the nuances of the rhetoric being used.
The court's decision reinforces the principle that witness statements should not become vehicles for argument or commentary. This trend has been criticized by judges in the past, who have decried the shift toward advocacy-style drafting. The use of AI makes it easier for officers or lawyers to produce these types of statements quickly, which can lead to a higher volume of low-quality evidence.
Moving forward, the judiciary will likely look for stricter adherence to factual reporting. The goal is to ensure that the court hears from the witness directly, rather than through a polished narrative constructed by a machine. This ensures that the sentencing and legal outcomes are based on reality rather than manufactured emotional impact.
What to watch for
Observers will be watching to see if the Court of Appeal's warning leads to formal new guidelines across all UK courts. While some coroners have already acted, a unified judicial standard would provide clearer boundaries for police and legal professionals. This may include specific prohibitions on using generative AI for any part of a witness statement.
There is also a growing conversation about how to handle 'hallucinated' citations in legal filings. As AI becomes more common in legal research, courts may need to develop standardized procedures for identifying and striking out non-existent authorities. This will be crucial for maintaining the reliability of the legal system.
Finally, the investigation into the officer's conduct will be a key indicator of how the police will be held accountable for AI misuse. It may lead to new internal training programs that specifically address the ethical boundaries of using tools like Microsoft Copilot in criminal investigations.
Try it
Sources and links
First reported by ukpolicelawblog.com. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.