AI Is Moving Deeper Into Venue Security Operations
Security Magazine opinion piece argues that large venues are shifting from passive surveillance toward AI systems built to detect anomalies, prioritize threats, and accelerate response across complex event environments.
A new opinion piece in Security Magazine puts the focus on how AI is being applied to physical security across stadiums, arenas, and other large event environments. Written by Ambient.ai CEO and co-founder Shikhar Shrestha, the article frames the challenge as one of scale: multi-venue events and high-density gatherings create too many moving parts for human teams alone to observe every risk in real time.
The core distinction in the piece is between legacy surveillance systems that record incidents and newer AI systems built to continuously observe, detect, assess, and support response. Shrestha describes that shift as a move from passive monitoring to proactive prevention, with AI analyzing camera and sensor feeds in context rather than simply flagging motion or storing footage for later review.
One of the clearest examples involves loitering detection. Instead of relying on fixed zones and pre-set dwell-time rules, he describes reasoning AI as evaluating behavior in context, including repeated circling near service entrances, coordinated movement across concourses, or unusual activity near restricted areas. In the stadium setting outlined, that can mean identifying a suspicious pattern near a VIP entrance after gates close and escalating a verified alert with visual context before a breach occurs.
Citing research that operators can miss up to 90% of activity after watching a single screen for 20 minutes, Shrestha argues that the issue is not competence but human limitation. The proposed model is one in which AI continuously processes every feed, triages anomalies by severity and context, and sends incident context with footage to responders, while human teams remain responsible for judgment and resolution.
Privacy is a central part of the argument as well. Shrestha says the most effective AI security systems should avoid facial recognition, should not build biometric databases, and should minimize retention of personally identifiable information. The article points instead to local processing and behavior-based threat detection, including examples such as identifying someone brandishing a firearm through behavioral signatures rather than relying on identity matching.
“Agentic physical security is not a future concept. It is operational now, and the organizations that embrace it will set the standard for how large-scale venues are protected.” – Shikhar Shrestha, CEO and Co-Founder of Ambient.ai
For venue operators, the piece reflects a broader direction in security technology: AI is increasingly being framed not just as a monitoring enhancement, but as a decision-support layer across ingress, perimeters, restricted areas, and real-time incident response. The argument here is not that human teams matter less. It is that dense crowds, distributed perimeters, and complex event footprints are pushing venues toward systems that can surface verified risks faster and help security personnel focus where human judgment is most valuable.
Shikhar Shrestha
