Thrift Library: Working with AI in Research and Writing Assignments: Faculty Learning Session: 4/12/24

Pros & Cons for Use of Generative AI in Research and Writing

Passcode: f3FPycC%

Pros	Cons
Involves algorithms that can generate text, images, videos, and music that resemble human-like creativity.	Need to have basic level of competence and subject matter expertise to create effective prompts and then edit results to produce reliable results. Must have base knowledge to compare output results.
Able to create new content or data (based on learned patterns) that is similar but not identical to existing data.	Does not pull sources from subscription databases - not able to penetrate those paywalls yet.
Good at summarizing content.	Difficult to cite - AI is not a person/author. Who is the author? Use organization as author?
Very good at creating templates (emails, job descriptions), lesson plans; recipes; training modules; workshop activities; brainstorming ideas.	Copyrighted information may be included (with or without appropriate permissions).
Use and evaluation of AI can encourage critical thinking - who is benefitting from AI and who isn't (discussion prompts).	Risk of plagiarism if the model reproduces sentences or paragraphs from its training data.
Library usage: Can be used to generate keywords, construct a search strategy, and recommend library databases for a topic.	Advanced options require a paid subscription.
	Can generate factually inaccurate or illogical answers (aka hallucinations). Need to be able to navigate the hallucinations.
	Training datasets are largely unknown to the user.
	May not include current information.
	Potential to amplify societal biases. How has the model been trained?
	Data privacy and security issues.