Final Exam - Intro to Gen AI Flashcards
3 characteristics of Gen AI models
ubiquitous
undetectable
transformative
ubiquitous
present, appearing, found everywhere
Characteristics of LLMs(large language models)
- super sophisticated at predicting text
- similar to working with a human who has been trained by humans
- good at some human tasks if you give them good instructions and depending on what tools they have available
Features of effective Gen AI prompts
- role and goal/identity, audience, purpose
- give very clear instructions (dont copy/paste)
- give examples and steps
Why should you provide role and goal/identity, audience, purpose
provides context from which to start
why should you provide clear instructions
improve effectiveness/relevence of output
why should you give examples and steps
provides better context
improves effectiveness of output
Ways to use AI
- where you are an expert and can assess quickly whether if AI is good or bad
- where using AI will keep you moving forward
- where AI is better than the best available human
- Where you need specific perspective to improve the work or when you need a second opinion
example of ‘where you are an expert and can assess quickly whether if AI is good or bad’
student researches topic on own first and knows required elements of proper citations and knows writing structure
example of ‘where using AI will keep you moving forward’
student is suffering from ‘writers block’
example of ‘where AI is better than the best available human’
student does not have a writing service appt. and missed drop in hours
example of ‘Where you need specific perspective to improve the work or when you need a second opinion’
student has trouble seeing the other side of the argument or cannot uncover underlying assumptions
ways to not use AI
- when you need to learn and synthesize new ideas or info
- when very high accuracy is required
- when the effort is the point. people need to struggle with a topic to succeed(productive struggle)
example of ‘when you need to learn and synthesize new ideas or info’
asking AI to summarize an article: shortcuts learning and breaches copyright.)
asking AI to create the components of an article
example of ‘when very high accuracy is required’
entering sources and evidence provided by ai without checking accuracy
using ai citations
example of ‘when the effort is the point. people need to struggle with a topic to succeed(productive struggle)’
submitting output generated by ai without applying the effort or adding value
not building necessary skills and competencies
