# Processing with AI ## Exploration : AI & Ethics Name: > Manoj > Subject: > Detect student presence in class using Face Recognition >[TOC] ## Design brief ### Biais If we don't source our dataset with enough rigor, the following biais might appear: >1. The recognition might work only for certain skin color or gender. >2. Faces in certain angles only might be recognized >3. There is a risk of incorrect recognition or mix & match of faces. We will ensure that our model is not biaised by: >1. Sourcing our data from a sample distribution that represent students from a mixture of skin tones. >2. Making sure our data take into account the orientation of image, lighting conditions and noise in the image. >3. We should perform image augmentation techniques so that the model will perform better irrespective of lighting conditions or orientation of a students face or minute changes in his facial features such has hair. ### Overfitting We will make sure our model does not overfit by > Checking the performance of our model on cross validation (cv) data, and also on cv data on which some sort of pertubations are performed. We divide the data into CV(16%),test(24%) and train(60%). We use this 16% of CV data and make sure the model is not overfitted. ### Misuse >We have to remind ourselves that our application could be misused by administrators who have access to facial data of students to do unethical morphing of images. ### Data leakage >In a catastrophic scenario, where all of our training dataset were stolen or recovered from our model, the risk would be that the images of students would be unethically misused specially for morphing of data and specific groups of students might become a target for racial/communal attacks also. ### Hacking > If someone found a way to "cheat" our model and make it make any prediction that it want instead of the real one, the risk would be that students would get proxy attendance while he might not be in the class, and the vice versa is also possible where a student might lose his attendance even if he/she is present in the class.