# Processing with AI ## Exploration : IA & Ethics Name: > Ethic issue for computer vision > Zhaoyi JI Subject: > Detect dermatological problems using Computer Vision >[TOC] ## Design brief ### Biais If we don't source our dataset with enough rigor, the following biais might appear: >1. A model that could provide 99% accuracy for men may make mistakes on women judgement. >2. A model that is designed for recognizing people's faces may miss black people's faces. >3. The information of minorities cases hasn't been collected so the model based on general information is not accurate enough. We will ensure that our model is not biaised by: >1. Sourcing our data from different races based on their proportion in patients. >2. Making sure our data take into account minorities cases in the overall patients. >3. Equally collecting data from the both genders. ### Overfitting We will make sure our model does not overfit by > Checking the accuracy of our model on random and unbiased data. > Train training dataset on test on validation dataset. ### Misuse >We have to remind ourselves that our application could be misused by organizations which pursue benefits illegally, to do behaviors that are harmful to patients' personal information. ### Data leakage >We have decided that our training dataset will be fully open-sourced, but before we made sure that all of our users have agreed on the public using of their information sharing on our platform, and the use of the information are legal. ### 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 a fake result will be shared to doctors and patients, which may influence people's understanding of the real disease and may bring people negative image of our model.