Policing in AI Summary
In case you didn’t get enough AI and policing in the last four posts, I had a fifth – MetaAI – evaluate the responses so you (and I) didn’t have to.
Here’s what we have:
Evaluation of AI Responses on Significant Developments in Policing
As a researcher in criminology with experience in policing, I’ll evaluate the responses from Grok, CoPilot, Gemini, and ChatGPT based on their ability to identify and analyze significant developments in policing from May 1 to August 1, 2025.
Criteria for Evaluation
- Relevance and accuracy of developments
- Depth of analysis and insight
- Clarity and coherence of writing
- Use of credible sources and citations
- Identification of potential second-order effects
Comparison of Responses
Grok
Grok’s response stands out for its clear structure and concise summaries. The three developments identified – Axon’s SkySwap Program, Police1’s “What Cops Want in 2025” survey, and the National Policing Digital Strategy 2025-2030 – are relevant and well-analyzed. Grok provides credible sources and citations, enhancing the credibility of its response. However, the analysis could benefit from more depth and insight into the implications of these developments.
CoPilot
CoPilot’s response is similar to Grok’s, identifying developments related to real-time crime centers, drone regulations, and AI-powered surveillance. While CoPilot’s analysis is clear and concise, it lacks the depth and insight provided by Grok. CoPilot’s response also feels more formulaic, with less attention to nuanced analysis.
Gemini
Gemini’s response takes a different approach, focusing on more structural and systemic changes in policing. The developments identified – the Supreme Court’s rejection of the “moment-of-threat” doctrine, the 12 Million Calls Initiative, and increased adoption of AI for non-facial recognition surveillance – are significant and well-analyzed. Gemini’s response is strong on insight and analysis, providing a nuanced understanding of the implications of these developments.
ChatGPT
ChatGPT’s response is notable for its identification of developments that are perhaps more provocative and attention-grabbing. The U.S. DOJ’s abandonment of federal consent decrees, Lincolnshire Police’s financial crisis, and technology-driven policing partnerships in India are all significant developments. However, ChatGPT’s analysis could benefit from more depth and insight, particularly in terms of potential second-order effects.
Ranking and Conclusion
Based on the evaluation criteria, I would rank the responses as follows ¹ ² ³:
- Gemini: 9/10 (strong analysis, insight, and clarity, with credible sources)
- Grok: 8.5/10 (clear structure, concise summaries, and credible sources, but could benefit from more depth)
- ChatGPT: 8/10 (identifies significant developments, but analysis could benefit from more depth and insight)
- CoPilot: 7.5/10 (clear and concise analysis, but lacks depth and nuance)
Overall, each response provides valuable insights into significant developments in policing. By considering multiple perspectives and analyses, professionals in the field can gain a more comprehensive understanding of the evolving policing landscape.
#themoreyouknow