Talk:AI and Recidvism

According to Google Docs, the length of this article is 855 words, which falls a little below the required range of 900-1100 words. The first paragraph summarizes the issue of recidivism and how AI may help predict and assess the likelihood of recidivism to mitigate its risk; however, it falls under the subtitle "COMPAS AI," so I would suggest the author move it up to a separate section to make it more apparent to readers. The body of the paragraph is clear and well-written. It explains why the prediction of AI is necessary, the advantage of AI over human beings, and the limitation of AI. One thing I think the author may improve is dividing each instance of AI into smaller subsections, like background, advantages, potential problems, etc., and using subtitles. The figures and statements are backed up by reliable sources, so it ensures the credibility of the information. The issue at stake is very clear -- the bias and limitations within AI and algorithms. The software cannot distinguish between correlation and causation, which will inevitably cause unfair and inaccurate predictions. Also, I think it will be a good idea to expand on the ethical concerns regarding bias, privacy, accountability, and control. The article is written objectively without expressing personal opinions or bias.