Facial recognition is a hotly debated topic as it impacts our personal lives. Everything from incorrect and biased labeling, to privacy concerns, to the fear of hacking and who owns your biometric data are all discussed when it comes to facial recognition. In the hands of the wrong people, facial recognition is insidious in its capabilities. Since Metaranx is a platform that democratizes artificial intelligence and therefore makes AI application creation accessible to everyone, we understand that there could be times when AI can be used for malicious purposes. In our terms of service, we will outline how we will handle this on a case-by-case basis, but we feel these cases will be rare as most applications of artificial intelligence support business functionality. Despite this, we believe facial recognition can truthfully only be used for malicious intent, especially by government bodies. We do not want to put the power of facial recognition in the hands of government bodies or private companies and have decided we will not allow it on our platform.
On the heels of corporate giants making similar decisions, we have decided to take a stance: We will not allow facial recognition tools or applications to be made, published, hosted, or integrated with on the Metaranx software platform or marketplace.
What is facial recognition artificial intelligence and why we do not allow it
A facial recognition system is a technology capable of identifying or verifying a person from a digital image or a video frame. Facial recognition uses biometrics to map a face in order to match an image or video of a person to other accessible information. For example, a person could use facial recognition to find all your social media and public profiles or a police department could run an image of you against their database. Currently, facial recognition is capable of 99.97% accuracy when matched to clear images, however in a crowded environment such as a sporting event (we imagine would be a similar environment to a protest), the accuracy ranged from 36% to 87% when applied to video. Imagine you were inaccurately identified by an algorithm with a 36% accuracy rate or even accurately identified. The conclusion to the application of this technology is not positive, especially in light of Black Lives Matter and the supporting demonstrations and protests. We want to prevent public misuse of facial recognition overall.
There are no laws governing tolerance for errors in facial recognition. Bias has an influence on outcomes, as well. One of Metaranx’s values and goals with allowing accessibility in artificial intelligence and democratizing this field is to eliminate bias. We don’t believe facial recognition helps us achieve those goals and that it actively goes against the values of our company.
Facial recognition used for good
Most of us are aware of how facial recognition is used by government bodies in countries to track citizens and manage criminal activity. There are a lot of negative applications for this technology. However, facial recognition isn’t all bad. Believe it or not, you use facial recognition or biometric data applications on a regular basis. Your phone or computer may allow you to access it through facial or fingerprint identification. Facebook may suggest people to tag in your photos, using their facial recognition technology. Voice recognition allows Siri to respond to only you. These features are common and even welcome in our daily technology. There are additional use cases with happy endings that prove facial recognition technology has positive capabilities.
Helping those who have addictions
Casinos have implemented facial recognition for their own personal security, but also to help with the voluntary exclusion list. In theory, when someone who has added themselves to the voluntary exclusion list enters a casino, facial recognition identifies them, which enables the casino to intercept this individual.
Finding missing persons
Being able to track and locate missing persons and children is a positive way that facial recognition can be used. It can move a missing persons case along faster to a hopefully positive resolution, as well. In India, facial recognition was able to find 3,000 missing children in four days.
Finding lost pets
Using facial recognition, companies like Finding Rover allow you to find your lost pet, source the owners of a lost pet you found, and even adopt. They have partnered with shelters and rescues to help get lost pets back to their homes. We will allow animals in image recognition so don’t worry, you can still build an amazing tool like this on Metaranx!
Healthcare and diagnoses
Some health diagnoses cause visual and observable symptoms on a patient’s face. Using facial recognition, those with these types of health problems can get a diagnosis faster. There are also interesting potential use cases for facial recognition in preventing burnout, reading emotions of patients in real time to deter aggressive situations, and accelerating identification to assist in diagnoses.
Other companies taking a stance on facial recognition
In the wake of the protests for George Floyd and Black Lives Matter, Microsoft, Amazon, and IBM have all denied police departments access to their facial recognition technology.
IBM is working towards tackling bias in artificial intelligence. They predict bias will only continue to increase over the next five years. IBM fully exited the facial recognition business citing racial bias and discrimintion. IBM’s CEO, Arvind Krishna has called on United States Congress to enact reforms to advance racial justice and combat systemic racism. IBM opposes and does not condone the use for this or other technology used for mass surveillance, racial profiling, or any other human rights violating purpose.
Microsoft will not allow the sale of controversial technology to police departments until there is a federal law regulating it. In July 2019, Microsoft addressed their steps going forward with corporate responsibility, laws and regulations, and privacy concerns surrounding facial recognition.
Amazon implemented a one-year hold on police departments using its facial “Rekognition” technology. They are also calling to governments to implement stronger regulations of this technology and the ethics behind using facial recognition.
Jordan Harrod discusses the stance these companies are taking on facial recognition over bias, misuse, and privacy concerns. Jordan goes into further detail about what these companies are doing with their stance.
<iframe width="100%" height="300" src="https://www.youtube.com/embed/1hkKvISIaEA" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
When these big players decided to exit, there were other companies who saw an opening in this lucrative space. Their departure may not mean an end to facial recognition for private businesses or government bodies but their firm stance is a great start and a strong message. We especially applaud IBM for taking a firm and long-standing position.
Allowing image recognition on Metaranx
We will still allow our creators and makers to build image recognition applications. There are so many great uses for businesses, personal, and fun with image recognition. In the early stages of our company, we will be reviewing the marketplace and what people are building on our platform regularly. We will also ask for support from our community to report any AI application on our platform that they feel is malicious in its intent. We understand that reviewing personally is not scalable in the long term but that’s what running an early stage company is all about. We will ask for your patience while we take on this challenge and for your help along the way.
Our stance on facial recognition will remain so long as we run this company
In light of Black Lives Matter demonstrations, we were envisioning the trouble a quickly-trained, ill-trained, or even successful and accurate facial recognition technology could cause. We were scared when thinking about a local police department using this type of technology to find people who participated in protests - pulling their image from CCTV footage, the videos and images posted publicly to social media, and news footage. We imagined false results and arrests and charges, but also far worse conclusions. We know there are already plenty of accessible facial recognition tools available. We won’t name any but there is one that’s fairly accurate and you don’t need so much as an account to use it - only a single image. Facial recognition is likely here to stay but we do not wish to play a role in it.
Facial recognition is an $8B a year industry and we know our decision to not support facial recognition technology on our platform may cost us potential investors and customers. We are prepared for that. While there are a few positive use cases for facial recognition, we don’t believe these positives outweigh the negatives, especially with no proper regulation in place. While you will still be able to create nearly any image recognition application you envision, you will not be able to build facial recognition applications on the Metaranx software platform. Ever.