I recently read an article about an artificial intelligence application that reportedly was able to predict future crime patterns with 90% accuracy; to say the least, I was intrigued.
These are very impressive numbers, and after six years of development of deep-learning models, statistical algorithms, and neural nets with the goal of identifying risk and stability levels and the leading indicators that could portent changes in those levels, I know that claims of 90% accuracy are pretty impressive. Digging a bit deeper, I found the claims were accurate, but predicting seven-day crime patterns in high-crime areas is much like predicting there will be fish in the ocean. So, what’s the real take away from this claim? Not much, actually. Why? Because it’s missing some key data to make strategic decisions around the predictions.
The Advantage of Predicting the Micro-Events
When I describe our PulseAI stability model, I often use Somalia and Sweden as examples. PulseAI identifies micro-changes in stability, looking for anomalies and trends across the globe that could put your people, places, investments, or interests at risk. To say that our neural net has learned that Somalia is unstable and Sweden is stable is not a revelation.
But, what the models are trained to do is look for the micro-events that could be leading indicators of disruption. By baselining normality in places as diverse as Stockholm and Mogadishu, the algorithms are able to detect anomalies that may normally go unseen and can provide indications of changing risk levels ahead of major disruptive events giving organizations decision advantage.
The ability to accurately identify timely potential issues before they happen is the best way to prevent catastrophes – by adjusting and pivoting in anticipation of the disruption. Efficient micro-event identification is a game-changer. This is the capability that is missing within the crime pattern predictions. It’s not enough to say that there will be crime.
This screenshot shows the Seerist PulseAI stability model over the past 60 days, preceding the recent bombings in Somalia. The data shows slightly continual decreasing levels of stability prior to the event.
Seerist Platform – Mogadishu – Decreasing Stability Levels Precede Recent Bombing
Three Limitations of Risk and Threat Assessment Solutions
This is where we have been, and before I talk about where we are going, let me give an admittedly simplistic description of the industry today to demonstrate why a change was needed. There were three significant gaps that other similar solutions couldn’t bridge, and this is why we formed Seerist.
- There are many very capable companies that can provide expertly written threat and risk assessments – yet their scale is limited, and they can be very expensive.
- There are also many data services that can give you a stream of news and social media – yet it’s not analyzed, often very noisy, and is being rapidly commoditized.
- There are very few companies looking at AI for security and threat analysis and not many with proven actionable insights.
We saw a hole in the industry and decided to fill it, making Seerist the only solution integrating the triad of Data + Experts + AI to enable accurate and actionable assessments on a global scale to help organizations save time, money, and lives.
The Power of Augmented Analytics
We understand that the true power of the Seerist solution is the combination of our artificial intelligence algorithms – that have been learning risk and stability patterns for five years – coupled with more than 10 years of data and analysis written by the world’s foremost security and geopolitical experts.
Seerist combines the veracity and depth of knowledge from experts with years of experience with the algorithmic speed and breadth developed by world-class data scientists. The result is assessments that go beyond a traditional stale report or streams of breaking event information to a truly immersive experience of in-depth analysis, real-time assessments, and forecasts of future risks.
The example below shows expert-written assessments, automatically generated breaking event reports, and AI-driven threat and risk forecasts looking at the recent tensions in and around Taiwan all in a single solution. The value is that analysts and users can get a much more comprehensive view into what’s happening and the intelligence to make more effective decisions.
Seerist Platform – Expert and AI-derived Analysis of China/Taiwan Tensions
The Future of Augmented Analytics
Over the coming months the focus is on creating a mutually beneficial relationship between the human experts and the machines where the experts are training the models and the models are providing insights to the experts.
The end result is an incredibly accurate assessment of threat, risk. and opportunities that users will be able to leverage to drive decisions and act with confidence.