Generative AI: Artificial Intelligence - Large Language Models

The Bad Side of Artificial Intelligence

While AI, in theory, is meant to be unbiased, it can only work with the data given, and that data reflects the unconscious bias of its creators as well as the AI creators. There are several instances in which bias has been on stark display in AI. One example of this is AI algorithms sending tech job openings to men but not women. 

There have been several studies and news articles written that have shown evidence of discriminatory outcomes due to bias in AI. Here are a few examples:

• Racial bias in a medical algorithm favors white patients over sicker black patients
• Statement of Concern About Predictive Policing by ACLU and 16 civil rights, racial justice, and technology organizations
• A review of predictive policing from the perspective of fairness
• Recent Advances in Algorithmic Biases and Fairness in Financial Services: A Survey

Ways in which AI can be used unethically

• Plagiarism 
• Spreading fake news
• Creating fake people to use in scams

Other Issues:

• ChatGPT and other A.I. bots often "hallucinate" facts and resources (and citations)
• Chatbots are not always great at answering questions accurately