Threat detected: Why the Department of Homeland Security is investing in AI 

The Department of Homeland Security is adding to its bench of artificial intelligence talent. Though many organizations are embracing AI for competitive advantage, the need within DHS may be different than what private-sector entities are looking for. 

The mission statement for the department is “With honor and integrity, we will safeguard the American People, our homeland, and our values.” Fulfilling this mission requires finding and securing information that could threaten the nation and its people from terrorist threats and actions. Whether such threats are foreign or domestic, the DHS seeks to deter, identify and apprehend such threats. Their success requires information on where such threats originate and how they will emerge to harm the nation and its people.  

This suggests that the DHS believes AI can play a role in collecting and analyzing such information. 

AI is superb in taking data — typically lots of data — and learning patterns that can uncover trends. In the case of Homeland Security, it can identify people, organizations or artifacts that pose a threat to the nation. 

Such data can uncover trends on how illicit drugs like fentanyl are coming into the nation or how and where sexual exploitation of children is occurring; provide more efficient immigration processing; improve airport security; and thwart cyberattacks like ransomware, to name just a few opportunities. 

The Transportation Security Administration is already embracing AI with its Credential Authentication Technology with facial recognition (CAT-2). This system ensures that all travelers presenting themselves at airport security checkpoints are eligible to fly that day and are who they claim to be. Such identity validation enhances the security of the air system by seamlessly facilitating secure air travel, while preventing bad actors with nefarious intent from entering the air system. 

AI can achieve with greater speed and accuracy what human-powered systems are unable to do. That is why the performance of X-ray technologies that scan cargo entering the country can be enhanced using AI algorithms. The same can be said to identify terrorists who work to conceal their identity and background to gain access and operate under-the-radar in the nation. 

Some will argue that using AI to penetrate the population is a violation of privacy, particularly when personal data are being collected and used. This is a common criticism of any AI-powered identity system like facial recognition. Such fears are misguided and misplaced. 

Data is oil for the AI engine. Securing such data is a top priority for DHS. Fortunately, most people and their data are not of interest to DHS. The few who are represent “needles in haystacks.” Finding such people is exceedingly challenging, and expensive. AI can make such processes more efficient and effective. The department also works to protect the majority of people’s privacy, not intrude on it, by keeping bad actors from accessing such information.

When a skilled surgeon performs a complex surgical procedure, the precision and sharpness of the scalpel targets the tissue that must be cut away while preserving healthy tissue. AI serves as such a tool, focusing attention on those who demand it, while leaving alone those with no background that requires scrutiny. 

Indeed, one can argue that the people who are most vociferous about privacy violations may also be those with the most to hide. 

AI has a definite place and value in homeland security. Without such a tool, near-peer adversaries will not only catch up to us but pass us, giving us a competitive disadvantage in protecting the nation. 

We live in a dangerous world. Ignoring such threats is risky. Not considering all possible tools available to thwart such threats is unwise. Assessing the appropriate use of such tools is appropriate and necessary.  

Investments in AI at all levels, including research, development and implementation, are critical to our nation’s well-being. They must also be prepared when new AI tools are needed to address unexpected and yet-to-be revealed threats, which are certain to occur at some point in the future. DHS is wise to attract the necessary AI talent and make appropriate investments that can help them uncover all possible avenues to protect the nation.  

Sheldon H. Jacobson, Ph.D., is a professor of computer science at the University of Illinois Urbana-Champaign.