There has been a flurry of noise around the use of facial recognition. Lawsuits, regulations, questionable-use cases and more. It was inevitable that much of this would end up in court, and it has. Now, finally, a case has been adjudicated and the verdict is in. This may be only one isolated case, in one area, but the implications are far-reaching. But before we discuss that, allow me to talk about privacy on a bit larger scale for a moment.
Privacy is a much bigger issue with the advancement of technology and it will be an even bigger issue once the digital transformation has occurred. Essentially, at that time, every second of our lives will be tied to some sort of digital interface and keeping the data connected to that interface secure will be a real-time, maximum effort – vigilance will be the watchword.
However, not all of this will be incumbent upon the individual entity, whatever it may be. Some of it will be incumbent upon others, some of it will be autonomous. But, what we do not have control of must have regulation – case in point, face recognition.
Circling back around to facial recognition, we are all aware of the controversy surrounding it. The looming question is whether or not obtaining facial images is a breach of an individual’s right to privacy, and permission must be granted before the image is captured and retained.
On one side of this are the privacy and human rights advocates. On the other side are security sectors, law enforcement, marketing agencies, private businesses and any number of other entities that can gain from knowing what you look like and what facial expressions can be interpreted as. It is no wonder these entities are pushing the limits, and privacy advocates are questioning their motives and rights to do this without guard bands.
There are many lawsuits, in various stages of progress, by any number of entities. Eventually, they will all be resolved, one way or the other. However, the one that is the most pertinent, IMHO, is in law enforcement. Finally, one of the suits has had a verdict.
The world’s first case against the controversial technology has been decided. This case, set in the United Kingdom, ended when judges dismissed the case brought by human rights campaign group, Liberty, on behalf a Cardiff resident whose face was scanned by South Wales Police during a trial of facial recognition. The scan was simply something that the police routinely do. And here, at least, the practice is now legal.
For law enforcement, facial recognition is going to be one of the top tools in the box. To be able to scan faces has so many positives, it is a no brainer – if it is done properly and the data is treated as intended. What that means is that if my face is scanned and there are no red flags, it goes into the bit bucket. It is not saved, archived, cataloged, categorized, or laughed at. Simply put, the scan goes away. I (and many people I discuss this with) are fine with that if that can be assured. Unfortunately, it cannot, thusly the need for bounds.
An argument from the plaintiff’s side is that the judgment does not reflect the very serious threat that facial recognition poses to our rights and freedoms. Facial recognition is a highly intrusive surveillance technology that allows the police to monitor and track us all.
That is, essentially, true. However, if the data is properly managed, no harm, no foul. If innocents are simply scanned and the data discarded, so what? And, if this catches a potential terrorist about to kill and maim scores of people, is there really any argument against that?
Many of these arguments are simply privacy catchalls around the protections afforded us by the constitution and other nation-specific documents. However, there are other issues that deserve discussion, which are more legitimate in terms of potential damage. One is accuracy, another is accountability. More than once, doubts have been raised about these two issues.
For example, in one of the latest trials, it was determined that the equipment used by the particular public safety agency recognized only about 19 percent of the faces on its watch list, correctly. That means four out of five people were wrongly identified.
While test results vary, the bottom line here is that some of it is relatively inaccurate, so there is merit to that claim. Other earlier tests have shown similar results. The fact is that today’s facial recognition software, especially fast, short, real-time scans has accuracy issues.
However, that is improving rapidly, especially with AI on the job. Experts predict that within two years, the accuracy will jump to 80 percent, or better.
Accuracy aside, there are compelling arguments for facial recognition in law enforcement. Other vectors, such as social media, private companies, and commercial use have different challenges, motivations and technologies and are a discussion for another time.
Facial recognition platforms offer a new paradigm for efficient and inexpensive (person-hours) large-scale individual identification. It is much quicker, more efficient and much less expensive than boots on the ground personnel.
On the other side of the coin, abuse is an issue. Uncontrolled, facial recognition has the potential for misuse just as much as any other platform. Unfortunately, even the police cannot be trusted 100 percent of the time to handle the data properly.
Therefore, restraints are necessary. However, like many emerging digital transformation platforms, it is a work in progress and we are still learning how to manage it. There will always be bad actors – some pushing the limits, others breaking them. For that reason, there needs to be regulation. How much is yet to be fully determined and it will need to be fluid, responding to evolving conditions.
For me, I have no problem with using technology, such as facial recognition, to increase the chances of law enforcement being able to derail crimes because they scanned a face that lead them to further investigation. I am willing to sacrifice a bit of my “inalienable” rights for a bit of alienable security.