COMM 1020: Introduction to Public Speaking

You've probably heard of ChatGPT, Midjourney, prompts and deepfakes. But what exactly is "generative AI"? That's what you'll learn in our explanatory video in two minutes! When new content is created by artificial intelligence (AI), it's called generative AI. This can be texts and images, but also videos, music or voices. Examples of generative AI are AI chatbots like ChatGPT and Google Bard, and text-to-image generators like Midjourney, Dall-E and Stable Diffusion.

There are several issues with AI sources' credibility. These include:

• Hallucinations: Generative AI has a tendency to create sources of information that do not actually exist. If you ask AI to write an annotated bibliography or paper for you, it will likely generate fake citations that look real. See an example of a fake citation from an annotated bibliography generated by ChatGPT below. 
  
• Unclear sources: Generative AI is trained on a variety of materials, like books, papers, and web content, and then uses these materials to generate responses. However, the response will not tell you what material it was trained on. This means that AI hides its information sources, making it difficult to determine their credibility.
   
• Bias: AI is biased according to the materials it was trained on and the choices made by the AI's creators. Consider what company or organization created the AI tool, and what their purpose was in creating and disseminating it, in order to evaluate any potential biases. 

While Generative AI tools, like ChatGPT, are notorious for hallucinating citations, some discovery tools that use AI are useful for seeking information, finding scholarly sources, and summarizing research. A few of the better-known options are listed below.

Review this guide by BGSU librarians Laura Sheets & Beth Fridrick for more information about artificial intelligence, including a glossary of terms, advice for prompt engineering, and a list of current generative AI tools.