This is the first post in a series on how artificial intelligence (AI) is impacting the physical security and business resilience space. In this post, we discuss how free or low-cost AI tools can save security operations teams hours of work every week.
AI's Impact So Far
Talk of AI continues to dominate leadership discussions across industries. A recent report by Morgan Stanley showed that mentions of “Operational Efficiency” are at all-time highs thanks to the adoption of AI tools.
For many functions, the use cases for AI were immediately obvious upon the launch of ChatGPT in 2022, and promised gains in efficiency are already materializing in areas like customer service where chatbots are doing the work of hundreds of customer service agents.
AI adoption in the physical security space has been far more uneven by comparison. First of all, physical security itself is a highly heterogeneous field that can have significantly different workflows and priorities depending on the company. The day-to-day of physical security at a hospital is very different from physical security at a manufacturing facility or executive protection team. That makes AI applications for security far more fragmented than applications than in other fields.
That being said, early adopters in the industry have already found quick and easy ways to save significant time using free or low-cost versions of popular AI tools.
One area in which AI - specifically large language models (LLMs) - excels is the summarization and synthesis of information. LLMs were trained by ingesting the entirety of the internet’s public-facing data, making them more knowledgeable than any individual on most topics. Furthermore, paid versions of tools like OpenAI’s ChatGPT can browse the internet for current information, which is critical for security professionals who need to react to current information. AI is a fantastic way to get caught up to speed on complex or unfamiliar topics in record time as well as sharing information to other team members or leadership in your organization.
Using AI to Produce Threat Reports
The best way to illustrate the power of AI in a security context might be to run through a demonstration.
Let’s say you’re a risk analyst for a large logistics corporation. Someone asks for the latest update on the conflict in the Red Sea where Houthi rebel attacks have been disrupting global trade, and they want the report submitted by end of day. Normally, this could take you all day to get done, but with AI, it can be done in minutes.
Start by going to your AI-powered chatbot of choice like OpenAI’s ChatGPT, Anthropic’s Claude, Googles’ Gemini, or Perplexity. (Note: as of June 2024, only paid versions of AI chatbots can browse the internet. Free versions still can offer a powerful way to quickly analyze and summarize data, but won’t be able to analyze real-time data, limiting their usefulness.)
After signing up, you’ll need to do what’s called a “prompt” (instructions for the AI to perform). In this example, we’ll ask the AI to create a report on the latest developments in the Red Sea where Houthi rebels have been attacking shipping vessels since 2023, significantly disrupting global trade.
“Provide an overview on the latest security developments on the conflict in the Red Sea and the current state of global shipping.”
Here, you can see the AI searches the internet for recent information and synthesizes it into a decent summary, even citing sources and providing links for independent verification.
However, there is still much to be desired from the output. The summary itself is a bit verbose and not particularly formatted, forcing you to do more to get it ready for presentation.
This time, you can try adjusting your prompt by specifying how you want the information formatted. Here, we will tell the AI to specifically format its report as:
Executive summary.
Timeline of the attacks.
Outlook for the shipping industry.
In fact, you can even instruct the AI to roleplay as a threat analyst to help coach how it should be thinking and writing. Here’s the result:
You can see that the AI produced a report that is well-formatted and easier to read. In addition, it linked multiple reputable sources such as the Council on Foreign Relations and industry-specific blogs.
You can go even further by asking the AI to reformat the timeline into a table for legibility in a presentation.
The Future of AI and Physical Security
As you can see, AI can be quite useful when it has access to the latest information. For most security professionals reacting to real-time security threats, paid AI tools with real time data will be worth it.
However, this is only scratching the surface of what AI can do for physical security teams. Customizing your AI assistant with information specific to your priorities can help unlock an even greater level of utility. For example, giving the AI context around specific assets you want to monitor as well as proprietary data from tools like Beakon can produce analysis and outcomes that otherwise would not be available.
If you'd like to learn more about real time threat data applications with AI, contact our team.
Comments