Join or Sign in

Register for your free asmag.com membership or if you are already a member,
sign in using your preferred method below.

To check your latest product inquiries, manage newsletter preference, update personal / company profile, or download member-exclusive reports, log in to your account now!
Login asmag.comMember Registration
https://www.asmag.com/showpost/34324.aspx
INSIGHTS

Generative AI in security: How video surveillance can benefit from it

Generative AI in security: How video surveillance can benefit from it
AI or artificial intelligence is not new in security. But how can security be impacted by generative AI? This article takes a look at generative AI in security and how video surveillance can benefit from it.
AI or artificial intelligence is not new in security. But how can security be impacted by generative AI, a branch of AI that can generate new contents based user requests? This article takes a look at generative AI in security and how video surveillance can benefit from it.
 
Generative AI is a type of AI that can create new content – be it text, images, video, audio or software code – based on user inquiries. For those of us having used ChatGPT to get detailed, article-length responses to an inputted query, generative AI shouldn’t be unfamiliar. Increasingly, Google is having this feature as well.
 

Generative AI in security

 
AI isn’t new in security. Security systems for example increasingly leverage the power of deep learning analytics to enhance object detection. But how can generative AI be used in security?
 
According to Will Knehr, Senior Manager of Information Assurance and Data Privacy at i-PRO Americas, generative AI can empower security by enabling systems to anticipate and simulate potential threats. “For example, it can create synthetic datasets to train systems for detecting rare scenarios like an intruder bypassing a fence or an unusual pattern of movement in restricted areas. Unlike traditional AI, which relies on historical data, generative AI can extrapolate new threat scenarios, helping organizations prepare for novel security challenges such as coordinated drone swarms or deep fake impersonations,” he said.
 

Applications in video surveillance

 
In fact, generative AI is already used in some aspect of video surveillance. For example, some cloud-based video management systems are already leveraging generative AI to enhance event search efficiency through intuitive natural language interfaces.
 
“Currently, generative AI is primarily used in management solutions like VSaaS. This technology allows users to effortlessly retrieve specific video footage by simply describing the scene in natural language. For example, a user could request, ‘Show me footage of a person wearing red clothing meeting with someone in the lobby around 3 p.m. yesterday.’ The system would then intelligently analyze the stored video data and deliver relevant results. It's like giving a task to an AI assistant,” Hanwha Vision said in its post about video surveillance trends for 2025.
 
But the benefits of generative AI in video surveillance go beyond typing a command asking the system to retrieve related video. According to Knehr, generative AI can help video security systems by simulating various scenarios that are hard to find in real-world data, such as people hiding objects or attempting to blend into a crowd while behaving suspiciously.
 
“For instance, it can generate video footage under different weather conditions, times of day, or camera perspectives, allowing models to adapt to unusual lighting or obstructions. This creates more reliable and accurate anomaly detection systems capable of identifying subtle threats like unauthorized access or suspicious loitering,” Knehr said.
 
Further, generative AI can do a good job detecting deep fakes, which contribute to the spreading of falsehoods and create security and privacy concerns.
 
“Generative AI excels at detecting deep fakes because it can be trained to recognize the artifacts and inconsistencies used in creating them. For example, a generative AI system might flag inconsistencies in shadow placements, blinking patterns, or lip-sync mismatches in a deep fake video. These capabilities are crucial in protecting systems where verification of video authenticity is critical, such as validating security footage or ensuring the integrity of remote monitoring,” Knehr said.
 

Future prospects

 
Looking ahead, generative AI is expected to have more video surveillance applications, moving from cloud to even edge devices.
 
“Edge AI cameras are predicted to evolve into intelligent AI agents capable of independently understanding and assessing situations, generating events, and providing real-time alerts,” the Hanwha Vision post said. “For example, intrusion detection will transcend the limitations of conventional systems by analyzing human behavioral patterns – such as running, loitering, or climbing fences – to determine intent, rather than simply detecting movement within predefined zones. Similarly, fire detection will move beyond merely detecting smoke or flames. By analyzing a broader context, including evacuation behavior and fire extinguisher use, these advanced systems will assess the likelihood of a fire and facilitate rapid response. This transformation will empower edge AI cameras to act autonomously on behalf of users, significantly enhancing security and efficiency.”
 
It adds: “Hanwha Vision anticipates that generative AI, currently limited to cloud-based solutions, will permeate both on-premise systems and edge AI cameras by 2025.”


Product Adopted:
Other
Subscribe to Newsletter
Stay updated with the latest trends and technologies in physical security

Share to: