This Billionaire’s AI Was Supposed To Speed Up Policing. It’s Not Going Well.
News Write-Up
Latest Development
San Mateo County, California spent around $12 million on C3 AI’s Project Sherlock.
Years after launch, there is no clear evidence that the system improved police efficiency or response times.
Figures and Data
$12 million: contract value with the county.
Zero published performance metrics: neither C3 AI nor county officials have released data showing faster case resolution or crime-fighting impact.
Background and Context
Project Sherlock was pitched as “AI to turbocharge police investigations,” promising to connect and analyze law-enforcement databases more quickly.
It even received early praise through a Smart Cities award, raising expectations of success.
In practice, however, the system failed to demonstrate measurable value, fueling doubts about costly AI deployments in sensitive public-safety domains.
Gaps and Missing Details
No clarity on root causes: Was the issue technical flaws in the algorithms, poor training, or weak integration with police operations?
No independent evaluation: external audits or reviews have not been shared.
Privacy and accountability concerns: the reporting doesn’t explain how citizen data is protected or how oversight is ensured when AI systems are deployed in policing.
Analytical Conclusion
This case illustrates a recurring problem with high-budget AI projects: big promises and millions in funding, but little real-world impact. For AI in law enforcement to succeed, it requires:
Transparent and measurable results,
Careful integration into existing workflows,
Independent oversight to safeguard privacy and rights.
Summary:
A $12M AI policing project by C3 AI failed to show results, raising doubts about costly AI solutions in law enforcement.