Social media is a world where trends are quickly placed. Many trends have rolled around our favorite platform in a nick of time. Wondering about how an app can suddenly become popular in the blink of an eye. Prisma was one among them. It was able to morph your photos into some sort of artwork which caught the fantasy of millions of users. And now it’s time for FaceApp and it has become a social media storm. If you’ve been on Instagram, Facebook, Twitter or on any other platforms, for the last few days you can’t have failed to notice the circulation of aged pictures of your friends and families. All of them have suddenly aged over 30 years from now. Well, what is the secret behind it? What is face app?? And most importantly why has it triggered concerns of privacy? Let’s find out.
What is FaceApp?
FaceApp is an AI-based photo editing app launched in 2017, developed by a Russian company named Wireless Lab. A free app with a single in-app purchase. By paying $3.99, we will get more additional specifications. Currently, it also ranks as the 4th grossing app on Google Play Store. According to sensor tower, since June 10 the app has been downloaded by more than 12.7 million new users.
At the time of launching it was only known for its ethnicity filters. Now, this app has been updated with a lot more features than before. It can convert your image into a younger or older version of yours by using filters. Converted images pose high level of reality which had made the app more trending and remarkable. It achieves this level of photo-realism by the assistance of AI. Curious to know how FaceApp makes such perfection? Continue reading.
Behind the Scenes of Faceapp
FaceApp was developed by a team of four engineers. They inculcated neural network to the former app for such photo realistic filters. In terms of a specific technology, they have used a “deep generative convolutional neural network” along with open-source libraries like Tensor Flow. Convolution neural network is the classification of deep neural network that deals with image processing.
Adding a simple smile or any other feature to a picture is a difficult task. But not any more with the usage of AI. With CNN any feature of an image can be easily processed helping us to get the desired results. With a perfectly labeled image these processes gets more easy. Image annotation is the technique used for labeling images.
Now, for instance take the above example of adding a smile. The face has to adjust the smile according to imagery. For that, we need to consider many factors like posture, lighting, skin color, the shape of lips and eyes, photo quality, etc. These features can be easily differentiated through a CNN. In addition to it, if we label the facial features more accurately and run through the deep learning algorithm, any effect can be made happen to the image. Various filters are available in the app and each are processed with different algorithms.
As we think, complete processing don’t happen in our handy devices. And this is where cloud computing comes into action. It is this feature of this app which has taken it into various controversies. Let us find out the reality behind it.
Is FaceApp a Threat?
As popularity increases eventually the questions regarding privacy were also increasing. One of the major questions raised about security concerns is that, whether the AI app could use the content uploaded by users. Will it use for “Commercial purposes” ?? Many press reports and social media posts raised these fears and doubts.
When every one of us is going behind it, has anyone bothered to read the terms and conditions of the app? Yes, that same lengthy document with lots of text that we usually simply skip by tapping on “I agree”. Read here to know more about the terms and conditions of the FaceApp. In the user content part of terms and services, they ask for ‘perpetual’, ‘irrevocable’ royalty-free and worldwide license to use the content. It means any photos uploaded to use any filter can be used by the company for promotional purposes anywhere in the world. And this single permission has taken the app into great controversies. The Russian origin of app is also considered another prime reason for these accusations.
While getting into reality, let’s think face app decides to use the photos they have collected through their user uploads. What are the possible misuse that they can do with that?? The first one can be to train algorithms for identifying your face. On a simple look this can never be your privacy thief. On considering the other way, Faceapp does not require users to log. There is no way to identify which photo belongs to whom, the needed info for persona recognition. And another important thing is while recognizing a person using a single photo, the accuracy can also be poor. Hence we can mark ourselves at a safe zone for now.
Need for such a huge database:
As mentioned above, such a huge database of image can be used for further training. And this is what the app builders can be focusing more on. No AI model is perfect and gains much added accuracy through continuous training and learning. Let us check out some of the features they might be interested in:
Face modification: FaceApp can use the photos collected for improvising themselves. The ability to change the age and specifications can only be achieved by training on tons of pictures. It would make sense for the company to improve its algorithms by feeding it more images. Using these images as a data set, they might come up with new features.
Face analysis: While Face recognition identifies the specific personality, face analysis analyzes the emotional responses, age, gender, etc from detected faces. FaceApp trained the app using open sourced databases and now can also use the collected databases.
Face detection: Similarly, face detection is about identifying human faces in digital images. We can train our model to recognize the faces in an image. For that, we first need a huge data set of faces itself. These systems could be built or upgraded with more face photos. Hence the photos collected from the FaceApp users might be used for this purpose.
Can we rely on it?
Responding to TechCrunch questions, Wireless Lab said that FaceApp has only access to the uploaded photos selected by users for editing. They are stored in a cloud because it is where it does all the processes. The cloud then retains the images for 48 hours only.
It is not only in the case of FaceApp. Are you sure that all your e-data are safe in other apps? Facebook says that you own your content. But the real fact is the company can use all our pictures along with all other actions we are doing in our account. We can end this agreement with Facebook by deleting it, but it is not ending. If you have shared it before it will follow you. It is not only in the case of Facebook, others like YouTube, IBM are also doing the same. People may worry about the threat of FaceApp using our image for their own purpose. But we have done it before. We are using many apps and services without even knowing its privacy policies.