What Does SEN Think Of Agentic AI’s Ability – We tend to think the security industry has been deploying agentic AI much longer than it realises.
Prior to the development of agentic AI, the electronic security industry sought ways to automate functionalities based on threshold breaches using networks of sensors, controllers, comms paths and control rooms. More recent solutions present situational awareness based on inputs and a range of other parameters – that also undertake automated processes.
There is a caveat. When it comes to speculations around what agentic AI might be able to do, something most security people discount is that electronic security solutions and the procedures and teams supporting them are life safety solutions.
This makes the reliability of agentic AI’s independent actions at key moments paramount for security professionals who are serious about their organisation’s security and safety stance.
It’s a fundamental that means agentic AI’s actioning of procedures must be carefully considered, especially now that the pace of AI development has accelerated to a dizzying degree.
It helps that the electronic security industry has been using early versions of non-agentic AI to parse data flows and handle intrusion detection and event classification and undertake automated responses far longer than most other industries have. This non-agentic AI is fundamentally hard coded – it’s not doing stuff on its own.
The industry’s history means that AI development – including a level of agentic AI development – is integrated into many electronic security solutions as emergent properties of software management systems.
When it comes to AI we love the idea of huge security, safety and automation systems that have fingertip awareness right down to sensor states and zone voltages and can act on this information.
Agentic AI could use sensor data flows to not only send reports or alerts on overall system states – many SMS and VMS systems do this already bear in mind – it could also trigger automation rules as thresholds are crossed while notifying system managers.
However, given experience with agentic AI, we know AI makes mistakes – not just little mistakes, either, but epic misconceptions that are irritating during trivial projects but could be catastrophic in a life safety system.
That means the rules and procedures built into any agentic AI’s operations need to be carefully considered and demand a degree of oversight, especially if the underpinnings of an enterprise system were undergoing remote updates.
Regardless, the potential of agentic AI across a range of use cases can’t be ignored, especially when it comes to parsing and responding to rivers of data and bringing key events to the attention of well-trained humans.
You can geek out on the Coasean Singularity of AI agents here or read more SEN news here. What Does SEN Think Of Agentic AI?
“What Does SEN Think Of Agentic AI’s Ability In Security Applications?”
Keep readingWhere this argument goes nextThe economics of AI agents — what happens to firms when the cost of coordination collapses — is the other half of this question, and it is being argued out in the economics literature rather than the security press.
What Does SEN Think Of Agentic AI - What Does SEN Think Of Agentic AI
