#79 - Quality Research Hasn't Changed With AI
In the age of AI, I feel like people’s abilities to pull credible sources of information is fading. Having talked to a lot of different people, society thinks that asking AI or getting on Social Media is considered “Research”. It is not. Academic Research as a discipline hasn’t really changed and the quality research that people need is often safeguarded or kept by institutions that require professionals to pay or contribute to their institutions. That’s the problem. There’s still:
Books that haven’t been scanned or put onto a public website for these large language models to scrape.
Websites like Reddit, that don’t allow traffic to go on their site and just collect it all for their AI models.
Many Research Institutes require pay walls to the PDF’s of their articles. So it is a bit of a mix and match. Yes, you can find good research by searching for free, but there’s still a lot that requires authentication or payment of some kind that you are a university professional or school/library professional
This is part of the reason why AI is going to struggle with quality. It’s pretty much already scraped all the free sources on the internet, but in new fields or fields that require a lot of quality research, people aren’t just going to fork it over for free so that Microsoft and Google and improve their AI models.
Social Media is even worse for research. A lot of the claims on there are not backed by any sort of science or any sort of knowledge base.
For example, I’m interested in gardening and so I go onto Social Media to find sources on Gardening. One Youtuber, says that you can make an excellent weed control spray by mixing soap with vinegar and a bunch of other different things. And yes, so viewers see this without further ado, and they go to their kitchens, make the spray, and start trying to control weeds with this. However, Vinegar as a substance might actually really alter the soil of your backyard and/or plants. Since it is so acidic. And so without any sort of backed up science to what you are doing, there’s no absolute way to really know that what you are doing is actually valid.
And it doesn’t even have to be published articles, we could stretch it and say that there at least has to be books published by experts in this field talking about this subject.
If you can at least find and purchase books on this subject, that increases the likelihood or weight that this information holds. It wasn’t just AI slop that has been thrown up by a Large Language Model and it’s not just some random person on social media trying to get views.
Anyways, those are my thoughts about the discipline of Research!
-Calvin
