Upside-down learning

Next: Connectivity

““In upside-down learning students create fairly advanced essays with LLMs.Grading is based on number of “hallucinations” i.e. errors they spot in AI generated textSpeeds up learning and fosters critical thinking of written text in general”“

One more for the education topic.

Local LLM allows also turning the learning experience on its head. As the LLM bot is excellent at answering even advanced topics and generating relevant text, this allows teachers to give earlier demanding tasks to students.

For example, to ask students in their teens to complete 30-40 page reports say in Aztec culture covering topics like political organization, religious beliefs and rituals, domestic life and marrying rules, social organization, ability and method of advancement in social ladder etc. LLMs are excellent at answering such detailed questions.

More advanced (older) students could be given bigger tasks like draft the history of the whole universe or history of the planet earth.

The task of students is to find errors in the LLM generated text material. The more errors found, the better grade. This type of learning allows students to cover fairly wide sets of topics fast while teaching critical thinking at the same time.

In broader terms this style can be seen as a way to teach students to think in terms of the big picture and use the computers to fill in the myriads of details. Then students verify the work of the computer by checking it and become themselves also familiar with the most important key factor. Computers are good at vast quantities of details and humans on creating patterns and driving strategies and this method guides students towards that.

Next: Connectivity