Reading for Sep 29 – Oct 6

Reading 1 “Artificial Intelligence’s White Guys Problem”

Through reading this article Kate Crawford explains how A.I can be very biased because it’s created and trained by humans. She talks about how technology can sometimes treat people very different due to their race or even gender due to the information that’s given to train AI can already contain those biases. An example is some AI systems have a lot of issues with darker skin illustrating that the technology doesn’t work  equally for everyone. Crawford’s main point is that AI isn’t automatically neutral or fair due to its computer system. The people who created the technology and the data they use can influence how it works. She also explains that these issues can have real effects on people especially on people who’ve already experienced discrimination. Overall, the article illustrates that we need to pay attention to who creates A.I, to what information it learns from, and even how it can affect different people instead of assuming that A.I will always make fair decisions.

Reading 2 “ChatGPT Is a Blurry JPEG of the Web”

From reading this Article, Ted Chiang explains how ChatGPT functions by comparing it to blurry JPEG pictures. When a picture becomes blurry you could understand what the picture is supposed to illustrate but you lose some parts of its original details. Chiang uses this comparison to explain that ChatGPT learns pattern from a huge amount of information and then uses those patterns to develop answers, but it doesn’t give the full or the exact complete information that it learned from. This is why ChatGPT can sometimes give answers that sounds very convincing even when wrong or even made up. Chiang questions whether ChatGPT understands what it’s saying or if it’s simply predicting what words should come next. He also explains that even though A.I can write things that sound intelligent, it doesn’t mean that it thinks or understands things the same way a person does. The article illustrates why people shouldn’t blindly trust A.I and should still check information and use their own judgement when using tools like ChatGPT.

Leave a Reply