4. COVID-19 is wreaking havoc around the world for the last couple of years. According to...

   

Transcribed Image Text:

4. COVID-19 is wreaking havoc around the world for the last couple of years. According to WHO, there have been around 350 million cases, including 5.6 million deaths worldwide. Rizwan became very distressed observing the pandemic. So, he started building an app that can predict if a user is infected or not. But how can he predict? Well, this is the era of machine learning, right? So, he built a machine learning model. But, he needs data to train the model which he got from the hospitals, taking no consent from the patients. Rizwan trained the model with this data and deployed the app publicly. So, now the app takes different COVID-19 related symptoms from the user as input and predicts if he/she is infected or not. Initially, the accuracy was very low. To solve the issue, he started taking other things e.g., occupation, address, financial status, gender, etc., from the user as input and trained the model with this newly added data. Note that he did not inform the user how this data is going to be used. At first, he got excited observing a better accuracy. But after a few weeks, he found the model predicts a positive infection always if the user is rich and lives in an urban area. This bias is introduced because most of the users are from the urban area and rich. But Rizwan actually built this app to help the poor in rural areas. Because there is a lack of testing facilities in the rural area. (a) Do you think the data collection process was ethical? If yes, explain. Otherwise, mention how is this unethical and what could have been done instead? (b) What type of bias is introduce here? Is this bias very harmful? What can he do to get rid of this bias? 4. COVID-19 is wreaking havoc around the world for the last couple of years. According to WHO, there have been around 350 million cases, including 5.6 million deaths worldwide. Rizwan became very distressed observing the pandemic. So, he started building an app that can predict if a user is infected or not. But how can he predict? Well, this is the era of machine learning, right? So, he built a machine learning model. But, he needs data to train the model which he got from the hospitals, taking no consent from the patients. Rizwan trained the model with this data and deployed the app publicly. So, now the app takes different COVID-19 related symptoms from the user as input and predicts if he/she is infected or not. Initially, the accuracy was very low. To solve the issue, he started taking other things e.g., occupation, address, financial status, gender, etc., from the user as input and trained the model with this newly added data. Note that he did not inform the user how this data is going to be used. At first, he got excited observing a better accuracy. But after a few weeks, he found the model predicts a positive infection always if the user is rich and lives in an urban area. This bias is introduced because most of the users are from the urban area and rich. But Rizwan actually built this app to help the poor in rural areas. Because there is a lack of testing facilities in the rural area. (a) Do you think the data collection process was ethical? If yes, explain. Otherwise, mention how is this unethical and what could have been done instead? (b) What type of bias is introduce here? Is this bias very harmful? What can he do to get rid of this bias?

Answer rating: 100% (QA)

a The data collection process in this situation was unethical Because Rizwan collected sensitive personal data from individuals without their approval ... View the full answer