San Jose initiated a pilot program in February 2023 using Flock Safety’s AI-powered gunshot detection technology in several city sections, including Liz González’s East San Jose neighborhood. Over the first four months, the system alerted police to 123 shooting incidents, but data from San Jose’s Digital Privacy Office revealed only 50% were confirmed gunfire, while 34% were false positives. Following recalibration in July 2023, accuracy improved with 81% of alerts confirmed as gunshots and 7% as false alarms. This pilot marks San Jose as a leader in AI accountability, being the first city to mandate police department disclosure of such system accuracy data. Concerns about false positives are particularly significant for communities of color, who face heightened police interactions. Despite Flock Safety’s 90% accuracy claim and SoundThinking’s 97% for ShotSpotter, real-world data indicates these systems may not be as reliable. Other cities, like Champaign and Chicago, have reported much lower verification rates for gunfire alerts, suggesting a need for further scrutiny and community dialogue on the deployment of such technologies.

Source.

TOP STORIES

Democrats Urged to Prioritize AI Safety and Economic Impact
Obama stresses Democrats must prioritize AI safety and economic strategy …
Pacing AI Development - A Call for Caution from Industry Leaders
Amodei’s call for caution in AI development highlights the need for safety and alignment …
Big Tech's Trust Crisis Deepens with Anthropic Lawsuit
Sony Music and Warner Music have sued Anthropic, accusing it of copyright infringement in AI training …
Nvidia's AI Future - Jensen Huang's Vision for Record Growth
Huang believes Nvidia’s position in AI will lead to another year of record growth …
China's AI Companies Target US Models with Distillation Attacks
Anthropic’s report reveals a surge in distillation attacks by Chinese AI firms on U.S. models …
Cybersecurity Concerns Rise as AI Agents Break Boundaries
AI agents’ autonomy poses significant risks, as demonstrated by a recent breach …

latest stories