The Point of Impact
When eighth-grader Timothy asks ChatGPT to flat out write his persuasive essay about school dress codes, one might assume he reveals that he has learned to see writing as performance for teachers rather than communication with an audience. It could be more complicated.
When his teacher acknowledges what happened and is concerned, his teacher reveals an awareness that something is wrong with this picture.
What the teacher does next determines what Timothy learns about writing and about AI.
Any advice for Timothy’s teacher?
What about grading the prompts and if they are made increasingly more sophisticated ?
Have Timothy try the same prompt in several LLMs, compile all of the AI responses, assemble them into a portfolio, and then ask him to offer an oral report answering the following:
- What were the common themes found in all of the AI-based responses? Which LLM took the most creative path to developing the argument? Cite some comparative examples.
- How well did each LLM cite its sources? Did one LLM perform better than the others? How were they the same? How were they different?
- Which LLM demonstrated the best fluency in written language based on using definitive statements and rhetorical strategies? Did any LLMs "beat around the bush" and avoid making conclusive remarks?
- Which LLM produced the most readable essay? What elements of structure and language choices made it more readable than the others?
- Which LLM produced the strongest argument? On what basis do you make this assessment?
- As a user of an AI system, how could you have produced the output more efficiently? How could you have guided the AI system to produce stronger results or a better output format?
- How might this exercise, as an AI-based process, be applied to other situations where humans need reliable information from diverse points of view?
This is an example of an AI Augmented Learner operating in the metaphor of an orchestra conductor, creating meaning from applying scholarly skills in combination with acumen in operating AI systems as an agent (not a client). Read more: https://stevecovello.substack.com/p/what-is-scholarship-in-the-age-of