If you have been looking to answer the question Is Facial Recognition Technology Safe, you are in the right place! Let’s start by looking at the basics of what facial recognition is, why it’s growing now more than ever before and why questions are being raised over its ethics.

What is facial recognition?

The basic concept of facial recognition is to apply machine-learning algorithms to analyze one or more images of human faces so that the system can “learn” the face. When it encounters a new image of a face, it compares features from this new image with those stored in its database and tries to find the most accurate match.

Early attempts at facial recognition technologies were used to identify individuals in crowds, but they quickly became more sophisticated. Today, they can be used to identify people not seen from afar. In the criminal world, the use of facial recognition technologies makes it possible to track people’s faces from a distance.

Facial recognition software can be used to identify people through images of their faces taken under various circumstances, for various purposes. This blog will explore the ethical implications of using facial recognition technology.

Ethical Issues

Privacy and security. Facial recognition technology can be used to threaten people’s privacy, as it may be used to identify individuals who are not aware that their images are being captured and analyzed. Law enforcement agencies sometimes find such surveillance useful for tracking criminals, although it has been the source of controversy as well as debate.

Facial recognition is a fascinating and innovative technology that has been used widely in criminal investigations and rehabilitation, as well as for commercial purposes such as marketing.

How do some facial recognition programs work under the hood?

• Facial recognition software works by analyzing images that precede the image that is being searched for. It usually uses three different methods to score images:
• A first method of analysis is to classify an image as either a person or not, based on algorithms trained upon training images. Images classified as “not people” are passed through further processing for “persons”. This process, called “supervised learning“, enables the system to identify persons by their features with little biometric knowledge previously gained about the subject’s face.
• A second method of analysis is to cluster images into groups, based on attributes that are obtained by counting the presence of specific features in each image. Each cluster in this method corresponds to an attribute profile. This process, called “unsupervised learning“, enables the system to identify persons by their features with little biometric knowledge previously gained about the subject’s face.

The main ethical concerns about facial recognition technology are:

• Privacy and freedom of expression.
  Photographing people in public spaces is generally legal, but biometric identification of individuals is not. On the other hand, facial recognition software was initially developed for security and law enforcement. It might be used against people who break into private spaces or against protesters demonstrating in public places.
• Data sharing and effectiveness of usage.
  The use of facial recognition technology has raised concerns about privacy and civil liberties. Companies such as Google and Facebook have created huge databases full of personal information about individual users which, if compromised, could compromise peoples’ privacy through impersonation or blackmail.
• Biometrics vs. biometry.
  Scientists are now able to extract more information about an individual, like their age, gender etc., than was previously possible with simple human portraits (e.g., photos). This opens the possibility that data mining (using complex algorithms) could be able to improve the accuracy of facial recognition software. It could also be used for discrimination and discrimination by proxy.
• Human biases in software development and usability testing.
  Facial recognition software can perform accurately on people who look similar (e.g., citizens of the same ethnicity), but it is far less accurate when applied to people who do not look similar (e.g., people with different ethnicities). This raises the possibility that facial recognition systems might be used selectively against (or for) certain groups of people, with the effect of widening existing inequalities.

Advantages of facial recognition technology

• As opposed to passwords and pin codes, this technology is more secure and less prone to human error.
• It also provides a solution to the problem of data breaches or loss of passwords, which can occur in a variety of situations, such as hacking or if a password is stolen.
• Facial recognition technology allows people to access their accounts anytime, anywhere.
• This technology can help government agencies organize large crowds and organize the flow of traffic.

Disadvantages of facial recognition technology

• Accuracy: A small fraction of images used in training datasets need to be retaken in order for the system to be accurate in all conditions.
• Time: Facial recognition systems can be slow, especially if they are designed with accuracy in mind.
• Privacy concerns: Since facial recognition systems can be applied to networked cameras, there are privacy concerns regarding the use of this technology.
• Accuracy: Facial recognition technology is still too inaccurate for full security purposes.

How reliable is facial recognition software?

It is highly effective in identifying people, provided it has enough pictures of an individual to identify them accurately. Each person’s face is unique, with approximately 80% of people’s faces being able to be uniquely identified by their features. If a person’s face is not unique enough, then one can use other characteristics such as height or clothes to help the software recognize them more easily.

Is Facial Recognition Technology Safe?

Facial recognition software can be turned off for security purposes. This is because the software only works with images taken in controlled environments where there are no obstructions to viewing the face, such as weather conditions. Also, other people in the image should not be wearing similar hats or other clothing that might cause confusion.

So, is facial recognition ethical?

It can be debated whether using facial recognition software is ethical since it violates people’s right to privacy.
Most faces in facial recognition systems are acquired from public spaces such as pubs or restaurants. This means that there is a chance that the face may not belong to the person whose face is being photographed and it can also expose people to security risks.

There has been a growing concern that law enforcement and national security agencies might use this technology in an unethical way. Such concerns come from: – The possibility of this technology being used to unethically monitor people’s activities or to monitor people for any reason.

The fact that facial recognition software is as accurate as it is now, and as the necessary algorithms develop better, this technology will become even more invasive and able to identify people even in low-quality images. This means that facial recognition technology is also open to abuse since it can be used in many ways which violate an individual’s right to privacy and freedom of expression. For example, a company renting out a private space could use this technology to monitor an individual’s activity when in that space.

In addition, facial recognition software might become a tool for unjust discrimination since it can be used in situations where people are judged for who they are rather than their actual abilities. For instance, if this technology was used in job interviews or professional spaces, it could be used to judge an individual according to their race or gender. This is because the technology by its nature uses stereotypes and algorithms to recognize a person’s face and these two can often lead to biased outcomes.

However, since no regulation exists on how they can use this technology, it is difficult to control the use of facial recognition software since there are no legal restrictions that can be put on them.

On the other hand, if facial recognition software was used in a way that did not violate an individual’s right to privacy, there are some advantages this technology can bring. Such advantages include:

• Facial recognition technology could be used to check the identity of criminals when they are in custody and when conducting interviews with victims and witnesses.
• It could also be used to monitor vehicles in parking lots. This is because most vehicle theft cases occur because of the use of duplicate keys. This can be solved by installing cameras at all key locations and this would require a new key to be programmed into the software, giving faces that match the one on the camera’s database access only.

With all this information, what’s your stand/take-away on the ethical nature of facial recognition?

Conclusion

Facial recognition is one of its kind, it has its own unique set of ethical concerns regarding privacy and freedom of expression, data sharing and effectiveness of usage etc. The key takeaway is that not all facial recognition technology is the same (just like not all cameras are the same), so you have to understand the pros and cons to choosing between various options.

Read an interesting article on the pros and cons of social media.