TLDR: Axon Enterprise’s new generative AI tool, “Draft One,” designed to auto-create police reports from body-worn camera audio, is facing significant criticism from civil rights organizations. Critics allege the tool is built to evade accountability for bias and mistakes by obscuring AI involvement and reducing human oversight. While Axon asserts its commitment to responsible AI and transparency, reports indicate that police departments are disabling crucial safeguards.
Axon Enterprise, a leading provider of police body cameras and law enforcement technology, has introduced “Draft One,” a generative artificial intelligence (AI) tool intended to streamline the creation of police reports. Launched in 2025, the software utilizes audio from body-worn cameras to automatically generate initial drafts of incident reports. However, this innovation has quickly become a focal point of controversy, drawing sharp criticism from civil rights advocates, including the Business & Human Rights Resource Centre, the Electronic Frontier Foundation (EFF), and the American Civil Liberties Union (ACLU).
Critics argue that “Draft One” is “allegedly designed to evade accountability for bias or mistakes.” A primary concern is the tool’s potential to obscure when AI is used in report generation, thereby reducing transparency and human oversight. The Electronic Frontier Foundation specifically highlighted that the system reportedly lacks version tracking and fails to clearly indicate which text segments were AI-generated, making it difficult to distinguish officer input from AI output. This opacity, they contend, makes it nearly impossible to hold either the officer or the AI accountable for errors or inherent biases.
Generative AI tools, including the custom variant of OpenAI’s ChatGPT used by Axon, are known for their propensity to exhibit racial and gender biases and to “hallucinate” or insert inaccuracies into texts. Civil rights groups warn that these tendencies could exacerbate existing inequalities within policing. The ACLU, in a white paper, explicitly advised police departments against using AI for drafting reports, citing concerns about unreliability, bias, lack of transparency in AI models, and absent privacy protections.
In response to these allegations, Axon Enterprise has stated that its product development is “driven by responsible innovation, grounded in a set of guiding principles.” The company emphasizes its commitment to ethical AI, aiming to revolutionize public safety while rigorously mitigating biases and risks. Axon asserts that “Draft One” is built with safeguards to ensure its use is visible at every stage, creating multiple layers of accountability and preserving the “crucial role of human decision-making” by keeping officers “in-the-loop.” They believe that clearly identifying AI assistance promotes accountability and public trust.
Despite Axon’s assurances, government documents obtained through freedom of information laws reveal a contradictory reality. Police departments utilizing “Draft One” have reportedly been disabling these very safeguards. This practice includes reducing or eliminating human oversight and deactivating features meant to prevent AI bias, making it challenging to audit AI-generated reports. For instance, a spreadsheet from the South Jordan Police Department showed the software generated over 900 reports for various incidents between September 2024 and April 2025. Similarly, the Fresno Police Department used the software for more than 3,000 incidents between December 2024 and April 2025.
Also Read:
- AI-Generated Suspect Likenesses: Police Practices Spark Public Inquiry
- Beyond the Hype: Unpacking the Overlooked Aspects of AI Accountability and Governance
This controversy is not Axon’s first encounter with ethical dilemmas; in 2022, most of its AI Ethics Board resigned over the company’s plans to integrate Tasers with drones. The current debate surrounding “Draft One” underscores the broader challenges of integrating advanced AI into sensitive public safety domains, particularly concerning accountability, transparency, and the potential for exacerbating existing systemic issues.
