The Central Blue is pleased to welcome first-time contributor Kristi Adam as she considers the historical developments and military implications of biometrics for the #ADFRAS2040 series.
Using biometric technology in security is not a new concept. Something unique to an individual – a fingerprint or a retina – has been used for identification alongside access cards and passwords for some time. Historically, biometric security technology has predominantly focused on facilities and network access and been conducted with the consent of the individual being identified. More recently, there have been significant developments in non-cooperative identification of individuals using biometrics. Using this newer technology, it could be possible to identify an adversary utilising extensive database and targeting systems automatically. While there are significant and promising advancements occurring in biometrics, it is also an ethical minefield. Due to these ethical considerations, it is unlikely that the Australian Defence Force (ADF) will progress to a fully autonomous human-out-of-the-loop targeting system. However, the ADF must understand how other military forces could exploit these technologies in the future.
Fingerprints were the beginning of biometric identification, but continually advancing technology has enabled significant developments in voice, gait, and facial recognition. Progression has been assisted by common, commercial-access systems like those used for accessing smartphones. As a consequence, there has been a significant increase in accuracy for facial recognition over the past four years, with the failure rate reducing from 4 per cent down to 0.2 per cent in certain circumstances such as passport photo identification. The United States Special Operations Command (SOCOM) have capitalised on this improvement with a project named Advanced Tactical Facial Recognition at a Distance Technology. This project has enabled a portable face recognition device that can operate at up to one kilometre. With the addition of thermal imaging, it is further possible for facial recognition to occur while a person is wearing a face mask, though with decreased accuracy.
Biometric identification technology has moved beyond just facial recognition; one example is the use of lasers to measure a person's cardiac signature at a distance of up to 200 metres. The system, also known as ‘Jetson’, was developed in response to another request from US SOCOM. Using laser vibrometry, the device measures the vibrations on the garment fabric sitting against the chest. The technique can be used to add additional unique data to biometric databases as individuals' hearts differ in size, shape, and contraction patterns. The main advantage of this type of identification is that, unlike facial or gait recognition, a heartbeat cannot be altered or disguised. Unfortunately, the technology is currently limited by the thickness of the garment worn by the target as the laser is unable to penetrate heavy fabrics or body armour. Further, it requires 30 consecutive seconds of scanning to make an identification, which means that the target would need to be sitting or standing still.
In contemplating how the ADF could utilise advanced biometric technology to gain a winning advantage, there are several operational concepts to explore. Using a layered approach, friendly forces could simultaneously conduct face, voice, and gait recognition, along with cardiac signature measurements to improve identification accuracy significantly. This data could further feed databases to either enable a positive identification or make connections from the continued use in one area. For example, it could assist in locating an individual from an earlier scan even when the identity of that person had not yet been confirmed.
The technology applications are very versatile - from humanitarian and disaster relief to base security. In the conduct of humanitarian missions, the technology could be used to locate missing persons in a natural disaster or war zone, while for base security applications, it offers additional safety through distancing personnel from the potential threat. Biometric technology software integrated onto Unmanned Aerial Systems (UAS) may be able to positively identify friend from foe, thus pushing boundaries out to circumvent attacks on-base infrastructure.
Then there are the potential combat uses. Automatic Target Recognition (ATR) has improved the efficiency of offensive weapon and surveillance systems. An armed UAS may potentially find, fix, and track an individual through biometric recognition. As with many other capabilities, the ADF will require access to common allied databases to maximise efficiencies and opportunity for success. The US has been building a database of biometric data from global operations, and access to this or similar allied databases could assist Australia in fully exploiting biometric capabilities.
Like many of the topics studied within the robotics and automated systems field, there are significant ethical questions to be considered. Regarding biometrics specifically, questions relate to the risk of misidentification as well as data protection. Biometric data collection and storage has raised privacy concerns globally. In China, there is an ongoing legal case concerning the collection of facial recognition data without specific consent. Chinese law has yet to catch up to rapidly advancing technology, with biometric data currently being included as personally identifiable information covered under non-binding guidelines within a broader data protection framework. This is not dissimilar to current Australian law, which has yet to move beyond fingerprint technology. To address concerns, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) is developing an Artificial Intelligence (AI) framework. The framework covers important topics regarding both data governance and automated decisions; however, its focus is domestic applications. In covering issues of privacy and democratic freedoms, the framework may serve as a useful foundation from which military applications can be further considered.
Although technologically possible, utilising AI reliant on biometric data to make life and death decisions and removing the human-out-of-the-loop for efficiency raises significant ethical questions worthy of deep consideration. Until ethical considerations are resolved, or at least further understood, the ADF will likely retain human-in-the-loop processes for targeting; however, the technology has other uses that will be worth exploring. The ADF must consider how and when other militaries could exploit this capability in the future, and what defensive capabilities could be used in response.
Pilot Officer Kristi Adam is in the Royal Australian Air Force and currently undertaking a Bachelor in Business through University of New South Wales in addition to a Bachelor in Global Security through Murdoch University. The opinions expressed are hers alone and do not represent the views of the Royal Australian Air Force, the Australian Defence Force, or the Australian Government.