Careful with your research

[Karlin]

I've run into a number of issues when researching. In my case, the context is Michelangelo, the artist. I'm mentioning these as an aid or warning to others. Any additional comments about this are welcome.
1. AI. If you Google a topic, you'll get an AI summary. In my case, I was checking out a shopping list that Michelangelo had writen (with explanatory drawings). AI suggested that the artist was iliterate. He wasn't. Apparently his servant was. It was clear what the AI engine had pasted together to come to that conclusion, which was amusing, but it illustrates the dangers of AI.
2. Modern secondary sources. They tend to interpret things in a modern way. I watched a documentary, which claimed that since Michelangelo was beaten as a child, it affected him psychologically his entire life. Maybe- but maybe not. Likewise for his nose that was broken- it was the subject of a modern psychological interpretation, that may or may not be correct. If you possibly can, read primary sources. Afterwards, do what you will with them, since we're talking about fantasy. But be aware of what the actual facts are.
3. Another example of modern interpretations. Michelangelo was likely what we would call today homosexual. Secondary sources often state this as a fact. The basis is poems he wrote to a young man (note #1- he was not illiterate), so it does seem likely. But what this meant in 16th century Italy may be quite different from what it means today in Western societies.

[As a personal aside- I'm aware that this is going to be a bit tricky for me to handle. I know a gay couple who live a traditional lifestyle (besides their sexuality). I should ask them about this, to gain some insight, but I'm not sure how open they are to discussing it. ]

Anyhow, the main point is to be careful when doing research. There's a natural tendency to glance at the first thing that pops up on the screen after a search. That is enough sometimes (when did coffee first arrive in Italy?), but can be misleading at times.

Karlin

 

[Rudolphthered]

The AI summary is tricky. It exposes you to information that may be misleading, especially when it takes platforms like Reddit as reference, where most content is opinion-based and often not research-quality.
The exposure to "fast food" information can also encourage complacency, making people less inclined to conduct deep searches of the topic at hand. Sure it may be easier and more comfortable, but it's a highly flawed way of learning in my opinion.

As a rule of thumb, always consult multiple sources and perspectives when researching in depth, and verify the references or bibliography.

 

[Gurkhal]

Yeah, using an AI summery as the only research seems like a bad idea. But are there people writing who actually do this level of shallow research?

 

[Rudolphthered]

I’m not sure about writing specifically, but at the institution where I study, some people use AI summaries to complete their assignments. They also use tools like ChatGPT, not only for APA style corrections but to develop brand new paragraphs.

Since we’re supposed to at least have some interest in the field—we’re earning a professional degree after all, I imagine some writers also end up doing shallow research.

 

[Gurkhal]

I’m not sure about writing specifically, but at the institution where I study, some people use AI summaries to complete their assignments. They also use tools like ChatGPT, not only for APA style corrections but to develop brand new paragraphs.

Since we’re supposed to at least have some interest in the field—we’re earning a professional degree after all, I imagine some writers also end up doing shallow research.

I am slightly shocked to hear this to be honest. Since I'm clocking in to be an old timer I will say that it was better in the "good old days" in this specific regard.

 

[CupofJoe]

Where I work, we know people are going to use LLMs to "help" their research, so they are asked to include the prompts they used. That way [it is hoped] we can see their critical thinking in action. Also as an extension to the plagiarism software that we also use, there is a feature that reverse engineers text. You can add in an essay and ask it to find out there the text came from.

 

[skip.knox]

Using only the AI summary for your research is not the fault of AI, it's the fault of the researcher. Historians are taught to consult all the sources, primary as well as secondary. It's not enough to know the primary sources, you need to know the interpretations and arguments derived from those sources. In many cases it's impossible to do this, or even impossible to know where all sources might be found, but that's the ideal.

Along the same lines, and more realistically, journalists are (or were) taught to not treat a source (somewhat different meaning here than in history) as reliable without two confirming sources. Translating this to Internet-based research, this means reading at least three articles or websites. And this being the modern Internet, it means looking carefully to ensure that Source 3 is not merely parroting Source 1. It's not the number of sources but their quality.

Of course, this all applies only to those writers who mean to have something realistic or based on reality. It's also fine to go with a legend or rumor as a basis for a story or scene. Just so long as we can tell the difference and don't mistake the one for the other.

Which is, after all, no more than what Karlin was advising.

 

[skip.knox]

Related to Karlin's point about secondary sources, I would draw a wide and bright line between scholarly secondary sources and YouTube. Even the videos done by scholars. The reason is simply that a video, even a hour-long video, adds up to barely half a scholarly paper and but a fraction of a monograph (academic speak for Significant Academic Book). It's the difference between a novel and a short story. If you are looking for depth or for nuance, you're pretty much stuck for a longish spell of reading. If you just want a quick intro, sure, video is fine.

And if you want to check a fact you heard in the video, check in more than one place.

Thanks, Karlin, for the post! (exclamation point dedicated to pmmg)

 

[pmmg]

I just avoid AI when I can. Google seems to like to do a summary, which for general ballpark stuff, is easy enough, but when I really want an answer, I just skip it and look for something else.

Course, AI will get better over time. Maybe more trustworthy. But...I avoid it on principle. I suspect it will never be trustworthy. Its a prisoner to its sources. It would be interesting to let AI run free and see what it concludes, but I suspect the powers that be will never let that happen.