AI Ethics (James)

This week we mainly talked about AI ethics. The rapid development of AI has brought many conveniences to people, but people will consider ethical issues when anything appears. Generally speaking, the ethical issues caused by AI include copyright, plagiarism, bias, etc. This week’s group study learned about the risks brought by the growth of AI language models. The growth of large language models will be biased against minority languages ​​such as Dhiveli. Because the expansion of language models is mainly based on the world’s common languages, minority languages ​​have a small base of use, which leads to bias in building AI language models. The courseera website pointed out that scholars can use the intersection of modern society’s political, social, and economic issues with artificial intelligence to eliminate the bias of AI language models. The time range of Chatgpt and Deepseek’s databases is relatively short. They came out in 2023 and 2025 respectively, so the scope of the database is relatively closed. For example, if Chatgpt is instructed to “find a few peer-reviewed articles”, the GPT database is limited and sometimes it will generate completely non-existent documents. If I can create an open artificial intelligence, I will try to collect all the literature in the 20th and 21st centuries to create an AI that helps locate and summarize literature (similar to perplexity)

source: https://www.coursera.org/articles/ai-ethics

Leave a Reply