Classified: How to (actually) research and not torture ChatGPT
Audio:
It is very interesting to me how students are so perplexed by the element or the reality of research. For me, it is the ultimate form of freedom from the syllabus, as it allows you to gather information or ideas on your own and relay them back to the teacher independently. However, yes, what I am seeing is that the student is copying and pasting the entire prompt into the AI search box and then proceeding from there. So, I am writing this post to help… mostly the AI, but still.
Consider AI a person, search engines, libraries
That is essentially it. The Artificial Intelligence models you guys are predominantly using are categorised as Language Processors, and guess what - we are kind of the same, aren’t we? The only significant difference is that they handle a million more bits of information than we do, and they are remarkably accommodating about it. So, so nice.
However, yes, you are getting a singular perspective if you dump the whole lot of information into their processors. Try instead to reform the prompt you are given, allowing the AI to not “answer” your question, but to perhaps interpret it differently, or even to allow you some angles for an answer, rather than the answer itself. This will undoubtedly help you get the resources you need from various areas online, rather than just from the trembling lips of your AI.
For example:
Instead of going - Create a new invention that will help your school. It has to be new and never done before. Show the links to the resources that inspired your invention.
Try this - Could I have some possible inventions that would help my school life? My school life is now fine, but not very fun. I would like some resources online for ideas too. It has to be novel.
See what I did? I defined the prompt further by attaching a definition of my school life, along with simply breaking down the prompt into its components (invention/resources/school/novel)
This way, you are actually engaging with AI as a tool, and not a slave - got it?!
Research via search engines, such as Google, is a traditional approach; however, simply copying the entire prompt will not yield satisfactory results. Try looking for the components of the prompt once again and Google the night away.
The trick to search engine research is not to look for the “perfect” resource, because you are never going to find it. A site with novel inventions for school life that have never been done before? Sure thing lol
Remember, they are libraries, so once you find a good enough resource, be inspired by its content and branch out from there. So, for example, if you stumble upon a site that has inventions that have already been thought of or made, add a small iteration to it based on your life and your school.
Ta da - you now have a resource to offer that “inspired” your invention.
Looking for multiple resources? Explore the various angles your prompt has to offer, not just the primary goal.
Look for angles, not answers
Which leads to this. Research involves not resources that circle the answer desired, but focuses on the process you took to get there.
So, it is crucial to break the prompt down into its various aspects and define them through research.
Going back to the invention example (true story by the way), here is a possible breakdown for four angles:
Invention types (possible options - maybe with two angles: one tech / one analog)
Invention types (already done - why they are successful)
School life (different issues students face and how tech can help)
School life before and after (how inventions helped build academic life)
There you go.
Allow the results to dictate and guide you
Do not hesitate to glean information from the sources or statements that either AI or an engine provides - that is, research beautiful people!
The best way you can do it is to dive in blind, but when I say that, I do not mean blindly (adverbs, people)
Again, an initial breakdown of the prompt is key. That will be the secret sauce to getting the information you need.
I remember another case where a student of mine was so perplexed by this prompt for an upcoming debate:
Are richer countries responsible for taking care of poorer ones?
Or something like that. The teacher just gave it to them with no context or materials to boot. Remember, do not see that as a failure (which it is), but an opportunity to be free in your research.
Sooooo - breakdown real quick:
richer countries - definition/examples by that definition
poor countries - definition/examples by the definition
take care = definition? financial? emotional? resource aiding?
case studies of when a “rich” country helped a “poor” country in these aspects above
pros and cons of it
your answer based on that
Ta da…
Tweak results to suit your narrative (responsibly)
Many times, I see students already have an idea for their answer, and so, get frustrated that the research does not align with what they are thinking of.
It makes me laugh because that is the antithesis of research, as research is supposed to inform you as you go along, and not the other way around. That would ultimately result in confirmation bias, which is not ideal. Or pretty.
However, if that is the road you would like to take, it is perfectly fine if the task is not information-sensitive, like the invention one I am currently using as an example.
Look at the source that is close to your intentions, and find components in it that match yours. Again, we are opting for a breakdown, rather than a perfect fit. With these similar elements, you can then say that it has an impact or influence on your work, which it can, or that it aligns most closely with yours.
Another good way is to not just break down the prompt, but also your thoughts. Get the components of your ideas and look them up! The results will closely align with what you are aiming for, with a degree of difference that would help in your further understanding of your own mindddd.
Start macro and dwindle down the search
This is essential in the thought process that links back to my initial point of not looking for answers. You are not. You are looking for avenues and roads that lead you to them. So, with the initial breakdown, do not look for a single, specific example, but widen the net:
Instead of - Which country helped another country with money?
Try - Examples of “first world” countries offering financial support to “third world” countries
I know you may think the differences are superficial there, but you have to respect the process. The brain will reward you with a bunch of endorphins because you actually did the task correctly, and not just scrambled for an answer and got a cheap dopamine hit after you pushed submit. Try it out :)
Any questions - feel free to ask below!
