What is Generative Artificial Intelligence (GAI)?

Generative AI is a type of artificial intelligence that can learn from and mimic large amounts of data to create content such as text, images, music, videos, code, and more, based on inputs or prompts.  (HUIT)

Generative AI applications use large language models (LLMs) that are trained on sets of data to interpret, organize, make meaning of and generate natural language (human like) outputs.   This type of artificial intelligence (AI) uses machine learning techniques to process large amounts of data, like text and images to identify patterns which are then used to create content and generate responses.


Some popular generative AI applications are: Chat GPT, Google's Gemini, Claude and image generators like DALL-E, Midjourney, and Adobe Firefly.

Many of these tools provide similar kinds of functionality, but each is distinct, using different LLMs, and providing different kinds of responses to the prompts you provide.

Depending on the outcome you are looking for, these tools can be very useful for enhancing your research.

This guide will highlight a variety of strategies and concerns for integrating generative AI tools into your research process.


Some Key Considerations

There are several considerations for using these kinds of technologies in your research. Below a few key things to keep in mind:

Commercial: Most of these tools are proprietary, which introduces concerns about how they were created, the exact nature of the data that was used to train them and how that data was processed.  As many of these tools capture and reuse user data/input, there are also notable concerns about privacy and the sharing of private information.   

Bias and Reliability: Limited transparency into the nature of training data used for generative AI products introduces signficant concerns about algorithmic, political, cultural and other forms of bias.  Additionally, these applications can and do produce false information (also known as hallucinations) which commonly appear to be plausible and realistic. 

Authority and Copyright:  These tools can not legally "author" content. Efforts should me made to cite their use for academic writing and research.