It is no secret that cognitive technology is changing countless industries worldwide, and in many cases, these changes have been objectively positive and constructive. However, as these technologies continue to become increasingly normalized aspect of our culture, the management of such technologies remains in flux. When it comes to security and public safety, some flaws to this growing technology can present a glaring risk to our general wellbeing.
Well intentioned as they may be, Artificial Intelligence (AI) and ML-based recognition tools, are creating a converse effect on the individuals and institutions they strive to protect. While I am not against the implementation of technology as a security force multiplier, they must be used to augment the human operator, and not serve as a replacement or excuse to hire less qualified (cheaper) personnel. This presents a long-running debate in security and crime prevention. For example, one of the most common security technology investments is in video surveillance equipment. While this is, without question, one of the most vital physical security investments an organization can make; it’s important to note that video surveillance is evidentiary in nature, thus recording crime vs. preventing it. If you buy cameras before investing in access control, alarms and the personnel to monitor the video and respond to potential threats, your investment may have been made in vain. This is an viewpoint I have held steadfast to for a long time, but recent developments and media stories have only reinforced its credibility with regards to the future of both security and public safety. Take, for example, recent developments in Metropolitan Police biometrics, in which the technology was found to be 98% inaccurate.
Finding a way around
Despite a variety of industry advancements in recent decades, there remains a sad but true concept within the security world: when a new safety concept emerges, criminals will try to find a way to sabotage it. This is best known as “the better mousetrap”. This notion is very true of recognition technology, which on the surface seems like it would be nearly impossible to penetrate in this manner. However, the technology’s continued flaws and loopholes actually make it more of a gamble within modern security protocol — especially when paired with the undying cunning of those striving to breach it. Even Microsoft, one of the world’s leading tech giants and a clear embracer of cognitive technology, has addressed the significant risks stemming from recognition software, and now, the company and its contemporaries have been left to scramble for ways to mitigate risk while finding answers to new emerging threats.
Leaving universal privacy concerns aside, countermeasures are currently being developed. In recent months alone, we have already seen hoods, disguises, large sunglasses and the emergence of 3D printed heads created to fool recognition technology otherwise assumed to be “impenetrable” when being developed by technologists without the appropriate “bad guys” to consult in red cell testing. While much of this phenomenon has been linked to mobile recognition security, in principle, it only opens another doorway for criminal exploitation. However, considering the allure of “lucrative contracts for emerging technology” rooted in recognition capabilities, it is safe to assume that this push-and-pull will remain a notable issue for a long time. Furthermore, behavioral detection programs deployed by agencies like the NYPD, Met Police (UK), and TSA have yielded a myriad of false alarms; mostly attributed to a lack of regular drills and response planning activities designed to act on alerts by such technologies.
Just like security protocol in general, we must not allow ourselves to succumb to distractions and wedge conversations and focus on what will work best for the vast majority, and with 2019 in sight, that does not currently include the presence of recognition technology. Therefore, it’s always important for organizations to have analog backups to all of their new technologies. Not only is this important for training the human element of a solution, but best serves the emergency management & continuity of operations side of the public safety equation, to assure that your security and safety operation is able to continue if and when technologies malfunction.
A. Benjamin Mannes, MA, CPP, CESP is a nationally-recognized subject matter expert in public safety and investigations. He is the former Director, Office of Investigations for the American Board of Internal Medicine, having previously served in both federal and municipal law enforcement & homeland security agencies and an established Public Safety Consultant. He has served on collegiate academic advisory boards as well as the Executive Board of InfraGard, the FBI-coordinated public-private partnership for infrastructure protection for over six years.