AI Aids Police Departments in Writing Report Drafts
In a bid to streamline the process of police report writing, Axon, a leading law enforcement tech company, has developed Draft One, an AI-powered software that creates the initial drafts of police reports. The software, which uses transcripts from body camera footage to generate draft reports, has been making waves in several police departments across the United States.
The software prompts officers to review and fill in additional details before submission, aiming to prevent errors or missing details. However, it's important to note that the original, AI-generated draft isn't saved once a report is submitted as final. This raises concerns about auditing narrative changes and tracking corrections.
One of the key concerns about Draft One is its potential for inaccuracy. AI systems, like Draft One, are prone to what is known as 'hallucinations'—making up or distorting facts—and these errors can be missed even during officer review. Some jurisdictions, such as King County, Washington, have outright rejected AI-generated narratives over concerns about untracked errors and unverifiable accuracy.
Another concern is the potential for embedded bias. Because AI language models are trained on vast internet data, they inherently reflect societal biases, including racial and gender prejudices. These biases can influence how facts are framed in reports, potentially affecting judgments on guilt or innocence.
Transparency and disclosure are also key issues. Many law enforcement agencies currently do not disclose when AI is used to draft reports, creating opacity in the investigative and judicial process. California lawmakers are advancing legislation requiring police to disclose AI use in report writing to enhance transparency.
Despite these concerns, Draft One offers potential benefits. It can significantly improve efficiency, reducing paperwork time by up to 70%. Officers remain responsible for final report accuracy, theoretically ensuring human oversight. However, unchecked reliance on AI risks undermining accountability and could introduce or perpetuate errors and systemic biases without proper safeguards.
In Fort Collins, Colorado, Officer Scott Brittingham has been using Draft One to write police reports, a task that previously took him 45 minutes but now takes just 10 minutes. The department of Fort Collins Police Department has made Draft One available to all officers, previously available to around 70 officers.
The use of Draft One by officers has reduced the time spent writing reports by nearly 70%. However, it's crucial to ensure that this time saved isn't used to increase the number of reports generated, potentially leading to an increase in bureaucracy.
Axon is committed to continuous collaboration with various stakeholders, including third-party academics, restorative justice advocates, and community leaders, to provide feedback on how to responsibly develop its technology and mitigate potential biases.
The rollout of Draft One comes amid broader concerns around AI in law enforcement, after experiments with facial recognition technology have led to wrongful arrests. The software uses bracketed fill-in-the-blanks to ensure officers read through the drafts to correct potential errors or add missing information.
In summary, AI-drafted police reports could streamline law enforcement documentation but carry risks of inaccuracy, embedded bias, and reduced transparency. Ongoing debates stress the need for strict human oversight, error tracking, bias mitigation, and clear disclosure to maintain justice system integrity.
Technology, such as AI-powered software like Draft One, is revolutionizing the process of police report writing by creating initial drafts of reports (first sentence). However, concerns about potential inaccuracies, embedded biases, and reduced transparency persist, emphasizing the need for continuous collaboration and strict human oversight (second sentence).
