In the article Ethics Act. Arm’s Length by Kate Crawford. Umm, which is an excerpt from her book Atlas of AI. She argues that researchers and technology Companies can become disconnected from the real people who are affected by the systems they create. In the article, Crawford argues that. When it comes to these AI system research, companies are more focused on the fact that it’s a tool, and they seem to overlook the ethical consequences of their work. Crawford argues that she believes that these AI systems are not supposed to be used in important factors of society, such as healthcare, education, policing, and anything that has to do with government benefits, because of the fact that they learn from data about real people and places like the police database sometimes has. Wrong information in their system Which can later lead to discrimination or historical biases based on the data that’s there. She argues that AI can produce or even strengthen these problems. I think Crawford’s main argument is the fact that AI lacks emotion and sometimes the things that AI is being used for, you need human emotions to connect to different people for these jobs for example, there’s been many deaths these last couple of years due to AI giving people bad advice because they started using AI for their therapy and these Ai’s lack emotion resulting a lot of people ended up committing suicide. Children were committing suicide because of the advice that they got from AI. When it comes to the medical field, especially a therapist, it needs that kind of ethical human emotion of sympathy, umm and AI, sometimes it’s brutally honest. It lacks empathy. I think what Crawford is trying to see is that they don’t need to get rid of AI per SE. They just need to be mindful of where the data that they’re getting comes from and whether it’s reliable. They need to figure out if people can be harmed because of these technologies that are being developed.
Another article that was provided was an article that was published by IBM which was called What are large language models? The article later explained that LLMS are a type of artificial intelligence that are trained to collect extremely large amounts of data. These AIS learn patterns in language and use those patterns to generate responses. For example, ChatGPT, Claude, Gemini. The article explains that LLMS generally work by predicting. A word or a? Of. That should come next. They use things called Transformers, which help AI to understand the relationship between different words and pieces of information. After reading this article, I think what scares me about LLM’s Is that these AIS can provide wrong information that they could make sound convincing, but they are incorrect. Similar to what Crawford was saying these AI systems can reflect bias. They can reproduce or amplify biases. Present in their training. From one article. You see that IBM is explaining the data language, how large it is, and how it’s trained. To perform. In January, Language and Crawford is asking us about the ethical side of where does that data is coming from and what happens to the people. Represented in it. Explains that LLM learned from massive amount of data, while Crawford shows that this data comes from real people One is focus on the technology behind it, where one is focused on the real people behind it, where the data is coming from the people before it gets to the technology. I think based on this article, IBM. Proved Crawford’s point in the sense where it tells you. The data is from a computer where Crawford is. Telling you that the data in the computer came from real people.
