Public safety is the top priority of most governments globally. National governments have historically embraced advances in technology to support public safety in a variety of ways, from electric lights used for safe streets to modernizing public health and law enforcement with emergency services. This pattern is continuing in the digital age, with technologies such as artificial intelligence (AI) and machine learning (ML) now being applied successfully to virtually every aspect of safety, security, and law enforcement investigations. Leading national security and law enforcement organizations worldwide are now applying AI to a wide spectrum of crime and threat challenges, including counter-terrorism, border control, human trafficking, drug trafficking, financial crimes, counterfeiting, extremist recruiting, and other law enforcement priorities.
Two main developments are driving the need for artificial intelligence and related technologies in public safety planning. The first is the exponential growth of publicly-available unstructured data, such as text, video, audio, and pictures and images. The amount of unstructured data available in the physical and cyber worlds today is astounding, and rapidly expanding daily. This means that much of the information that could be potentially helpful in an investigation or in public safety planning already exists. Somewhere in the vast, amorphous ocean of cyber space, there are most likely hundreds of social media posts, pictures, videos, and other types of information that would help investigators to uncover crime rings or help investigators and analysts counter and thwart a terrorist attack. The problems lie in the sheer amount of information – usually enough to require a team of experienced researchers countless hours to sort through — and the lack of tools to show apparent connections between related data in an automated manner. Fortunately, AI has the ability to scour all available digital sources and make connections that the best human analysts simply do not have the time or resources to identify. The best of these AI systems are also developed to alert analysts and investigators to significant leads, connections, and newly-acquired relevant data.
Another development driving the use of AI in public safety efforts is the expansion of Machine Learning, which includes the subsets of Artificial Neural Networks and Deep Learning. These revolutionary technologies have taken AI to new heights, allowing users access to image and pattern recognition, machine translation, and natural language processing that would have been nothing more than science fiction just a decade ago. In an age of a more global economy and transnational crime, the ability to analyze information in multiple languages, dialects, and cultural contexts is absolutely crucial. Fortunately, the technologies have kept up with the demands of the day. When applied to investigations, simulations, projects, and operational plans, these systems reveal hidden layers of reality that can be vital to providing key insights.
Here are a few useful applications that demonstrate how AI and analytics can help with public safety.
Application 1: Quickly Analyzing Massive Numbers of Documents
One of the simplest but most helpful applications of AI is to comb through thousands of pages of written documents, such as social media warrant returns, to locate references, keywords, names, dates, locations, social security numbers, and other salient bits of unstructured information that can be linked, analyzed, and presented in an understandable, user-friendly format. The better systems can carry out these functions in more than 100 different foreign languages, factoring in regional dialects, slang, and abbreviations. The uses of this kind of analysis are virtually limitless for identifying relationships, establishing opportunity and motive, forecasting future threats and criminal activities, and identifying vulnerabilities.
Application 2: Analyzing and Leveraging Massive Numbers of Images
Similar to the AI ability to analyze large amounts of unstructured text in documents, AI solutions can also be used to find, sort, classify, link, and analyze thousands of images from social media, databases, and many other publicly available and proprietary platforms and sources. There are many high-impact automated results, including identifying structural vulnerabilities through aerial or geospatial images, matching images of suspects from security cameras with photos in social media or various databases, identifying posted locations of persons of interest by flagging identifying background markers like signs or distinct geological features, and a broad range of other applications. In some investigations, AI is critical to finding images of stolen property being advertised for sale on the “dark web,” or being displayed on social media. This can also be applied to flags, insignias, and other visuals used by terrorist organizations and other extremist groups, and to a range of biometrics that may have been recorded in databases.
Application 3: Highlighting Hidden Connections
In addition to the analysis of text and images, there are many other ways in which AI systems bring insights that would be virtually impossible otherwise. Combining textual analysis and image analysis with other types of databases, biometrics, criminal and medical records (both public and proprietary or classified, depending on the user’s access), social media platforms, financial data, purchasing records, travel records, and other information, these systems can highlight connections, networks, patterns, and activities that otherwise may have been missed.
Application 4: Ensuring Safety of Public Figures
Public figures like politicians, diplomats, military leaders, and corporate executives are always at risk of being attacked. This brings significant challenges to those who are tasked with safeguarding them. Fortunately, security professionals now have new solutions and capabilities to collect and analyze threat indicators and alert them to potential dangers.
The cognitive computer vision and pattern recognition capabilities of AI are capable of interpreting the behavior of suspicious individuals through security camera footage and other sources of video, flagging indicators of harmful intent. People in an area who appear to be hiding weapons, placing explosives, or doing anything else that could be potentially dangerous, are brought to the attention of authorities for further investigation. With this real-time analysis, agents and analysts are able to respond immediately to potential threats.
Additionally, these systems can quickly match their visual data to any publicly visible cyber connections of the person being observed. If this individual has identifiable links to criminal groups, extremist organizations, or criminal activities, authorities can be made aware of this immediately.
Application 5: Highlighting Connections to Terrorist Groups
As violent extremist ideologies spread ever more easily through digital communications, the membership in terrorist organizations continues to grow globally. As international travel, commerce, and communications become faster and easier, extremist involvement has become less regional and less predictable. This means that it is easier than ever for members of these groups to infiltrate governments, militaries, and private sector organizations for various nefarious purposes. This is why AI based on open-source collections and analysis can be easily and effectively be applied to background investigations that include broad, multi-platform searches to find any connections to violent groups. Through the methods mentioned above, including analysis of images, text, and records, AI systems can alert users to a person’s link to a terrorist group, even if steps were taken to obfuscate that connection.
The manner in which AI can be used to improve public safety can be applied to airports, train travel, communications and power infrastructure, hospital backup systems, water supplies, disease prevention, and virtually any system or operation that affects the well-being of a population. As the technology continues to develop, even more applications are likely to emerge, and more lives may be saved.