# Processing with AI
## Exploration : 👩‍⚖️ Ethics of AI

Name:
> Patrick Billiottet
>
Subject:
> Monitor student attention in class using Computer Vision

>[TOC]

## Design brief

### Bias

If we don't source our dataset with enough rigor, the following bias might appear:

>1. Wrongful cheating recognition
>2. Discrimination
>3. Unfairness

We will ensure that our model is not biased by:

>1. Sourcing our data from as many datasets of students cheating and diverse demographics.
>2. Making sure our data take into account each student action independently. Not because "that" student has cheated before that he is morelikely to cheat again. Be fair to all students.

### Overfitting

We will make sure our model does not overfit by regularization. Regularization refers to a broad range of techniques for artificially forcing your model to be simpler.

### Misuse

>We have to remind ourselves that our application could be misused by teachers who don't like certain students to make sure these student fails, or even worse, pedophiles, who could enjoy the sight of young students under stress in a classroom, during an exam for example.

### Data leakage

>We have decided that our training dataset will be fully open-sourced, so other companies could use it, and notn only schools. But before, we made sure that before being used in a certain workplace, the employer would have to input the number of employees he wants to observe, and each employee has to sign a consent form. Like so, spying on unaware employees won't be possible. It's a precaution on privacy.

### 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 cheaters, or dishonest workers could get away with it. And in doing so may also jeopardize the rest of the students or workers who are innocent.