# Processing with AI ## Exploration : 👩‍⚖️ Ethics of AI Name: > Zixin ZHANG > Subject: > 🔬 Detect rare diseases in medical results (blood tests, MRI, radiology, etc.) using Machine Learning >[TOC] ## Design brief ### Bias If we don't source our dataset with enough rigor, the following bias might appear: >1. We may misjudge new patients' situation >2. We may use the incorrect medicine in new patient >3. We may overlook a possible new disease >4. We may misjudge the severity of the condition We will ensure that our model is not biased by: >1. Sourcing our data from abundant patients from the hospital from various countries >2. Making sure our data take into account the case that have similar similar symptoms but different pathologies >3. Thinking of different sector, like genger, age, nationality, use these sector to conclude the patients more detailed, to define the disease more accurately ### Overfitting We will make sure our model does not overfit by > - Checking the accuracy of our model on previous patients > - Testing experiments in multiple cities and countries ### Misuse >We have to remind ourselves that our application could be misused by >- Third-party medical companies to conduct pathology experiments >- Insurance company to gets patients' specific information >- Pharmaceutical companies to contact patients privately in order to selling specidic drugs ### Data leakage ***Choose** the most relevant solution concerning your **dataset** privacy:* The data is containing privacy and abundant data from rare disease petients, which need to be the closed source. It's not suitable to open to the public as it infringe on privacy and cause social discrimination. >**🔐 Closed source:** In a catastrophic scenario, where all of our training dataset were stolen or recovered from our model, the risk would be >- The private information of all patients will be open to the third-party medical companies or pharmaceutical, which will be misused to do test their product nonstandard. >- The MRI, radiology,pathological report and related therapeutic process will be open to the public, which may cause the large-scale research in, the results from underqualified institution will incorrectly guide public opinion. ### 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 > - The patient's life will be threatened > - The patient may therefore receive a completely different treatment > - Wrong medication can cause serious sequelae > - Medical research on such conditions will be impaired and will hinder human breakthroughs in this type of rare diseases > - The reputation of doctors and hospitals will be severely criticized by public opinion > - Doctor-patient relationship deteriorated further