Assignment 6

I played around with Quick, Draw! and I realized how well the program recognizes a lot of doodles that were drawn terribly/effortlessly. It seems that many individuals have drawn similarly terrible doodles since the model almost immediately recognized my doodles.

• Does the dataset contain data that might be considered confidential?

I don’t think that is the case since most of the doodles are very rough due to the limitations on how you can draw. Most of the people must have used their laptop mouse or trackpad, which aren’t the best tools for accurate drawings. The drawings in the dataset do not contain anything that would reveal something about a particular person.

• Does the dataset contain data that, if viewed directly, might be offensive, insulting, threatening, or might otherwise cause anxiety?

Though it doesn’t affect me, a data that portrays something related to religion, culture, tradition, or identity in a negative manner would be considered offensive. For example, the data for The Great Wall of China could contain drawings that are drawn carelessly or offensively that would undermine the value it has for people living in China. So the data can potentially be offensive or insulting or threatening if it’s on a sensitive subject. Images are ultimately subjective and can potentially display the artist’s perspective.

• Does the dataset identify any subpopulations of people (e.g., by age, gender)?

The data for the lipstick would pertain to many women whereas data for cars would pertain to many men. So it really depends on the subject.

• Were individuals (e.g. players of Quick, Draw!) notified about the data collection?

Before starting the doodle, the website has a clear message conveying that by drawing, I would help teach the neural network by contributing to the world’s doodling dataset. Thus, anyone who starts the doodle would are notified that their drawings would be available to the public as well as the developers.

• Did these individuals in question consent to the collection and use of their data?

By clicking on “Let’s Draw!” button below the consent message, the individuals are giving consent to the collection and use of their data.

• If consent was obtained, were the consenting individuals provided with a mechanism to revoke their consent in the future or for certain uses?

No, it doesn’t seem like there is a ‘back’ button to undo a doodle or revoke the consent after completing 6 doodles. It only says how many doodles the network was able to identify and gives you an option to share the drawings.

• Are there tasks for which the dataset should not be used?

I don’t think it should be used for anything that categorizes people based on the different drawings submitted. It should not be used to compare it to the real life object.