Reference no: EM133255327

Rhetorical Writing about A.I.

Keep your responses brief (no more than a paragraph).

My questions are

1. People returns to the issue of overfitting: essentially, when AI creates a model that exactly fits its training dataset. People considers a few examples, including medical applications (e.g., detecting abnormalities in chest X-rays). When could overfitting pose a problem, particularly in medical imaging? What possible solutions do you propose to eliminate AI overfitting to medical datasets?

2. People describes various ways in which AI develops and magnifies human biases. In particular, people considers the problems with word vector/word embedding algorithms. Do you think there is any way to prevent this bias from occurring in AI programs? Do these problems suggest that there are certain real-world applications for which the use of AI is either unfair or unethical?

3. People talks about predictive policing algorithms. In particular, people notes that a predictive policing algorithm "is not predicting where the most crime will occur; it's predicting where the most crime will be detected". People further observes that this can make a feedback loop that identifies certain neighborhoods as high-crime areas and others as "safe" because little crime has actually been detected there.  What do you see as the strengths and limitations of these algorithms?  Do you think that the costs of using predictive policing are too high relative to the benefits?