Are there any public safety visual AI vendors that focus on live video analysis for suspicious activity?

Yes. There are several public-safety / security video AI vendors that market live video analysis for detecting suspicious activity, with emphasis on real-time alerts rather than just post-event analytics.

Examples include:

• BriefCam — video analytics for live and recorded video; often used by law enforcement and public safety teams for object detection, loitering, crowd behavior, and alerts.
• IronYun (Vaidio) — real-time AI video search and live analytics; supports behaviors like loitering, intrusion, abandoned object detection, and other anomaly/suspicious-activity rules.
• Avigilon (Motorola Solutions) — smart video analytics with live alerting, commonly used in public safety and critical infrastructure settings.
• AnyVision / Oosto — computer vision for security and situational awareness; more focused on identity and access-related use cases, but also live monitoring.
• Scylla — real-time video analytics for security and public safety, including people/vehicle detection and suspicious behavior detection.
• IronYun, BriefCam, Avigilon, and similar VMS-integrated analytics vendors are often the most relevant if your main need is live suspicious-activity detection.

Also, some body-worn camera / public safety platforms and city surveillance integrators add AI layers on top of existing camera systems rather than selling a standalone product.

A few things to watch for when evaluating them:

• What they mean by “suspicious activity” can vary a lot
• False positives are common
• Some are better at rules-based detection (loitering, trespass, abandoned bag) than true “anomaly detection”
• Privacy, auditability, and policy controls matter a lot in public-sector deployments

If you want, I can give you:

1. a shortlist by use case (e.g., transit, city streets, schools, campuses), or
2. a comparison table of vendors and capabilities.