Transcription
Most researchers use AI wrong. It's not about which AI tool you use, it's about how you use it. It's about building a system that allows you to really skyrocket your research paper output. That's why in this video I'm going to show you the four levels that such an AI research system must have. And if you follow them, then by the end of this video you will be more effective and you'll be able to publish more papers and you'll use AI better than 95% of researchers out there. So let's dive right in.
Now, the first thing that we need to clarify before I show you that exact system and framework is that it's not about which AI research tool you use. Like probably one of the most frequent questions that I get from our clients on published researcher or in the comments section to this YouTube channel is like, "Marek, what is the best AI tool that I should be using right now? Or you've done a video on JenAI, can you please do a video on this other tool that just came out? I'm sure it's better." 80% of your success comes from how you actually prompt AI, how you interact with it, the framework and the layer of tools that you build for yourself to achieve your bigger goals. So now that we have that out of the way, there are basically four levels to a proper AI research system that will allow you to skyrocket your research paper output.
Now, the first one is the knowledge [music] base. Because the days of like you as a researcher having to comb through the literature manually, get insights manually, read hundreds and hundreds and hundreds of papers every single year are frankly over. Where your brain power should be going is connecting the dots and really working on high-level problems rather than, you know, doing the manual work because AI can do it much faster. So the first part of that system is the brain or the knowledge base, if you will. And I think the best AI tool right now for that is probably Notebook LM by Google.
So, the way it works is that you can create notebooks in here, right? And then you will upload your papers to those notebooks, or you can also do deep search to add more sources in here, right? And it will if you switch to to deep search, it will scour the web to find answers to those questions, right? Or you can also use your Google Drive to upload the papers. Obviously, for, you know, academic work, the best thing to do is just to upload those papers. And then you can think of this as your brain, essentially. And the processing power of Notebook LM is really superior to a lot of other AI models. You can really upload up to 300 papers [music] on the paid version of it, and it will hold that context when answering [music] your questions, right? And in the answers, it always bases them on the papers, right? And it even gives you the extract in here. So, whatever question you ask it, right? It will always be based on the facts, okay? Which I think is great. Like, it cannot really hallucinate. I mean, yeah, it probably could, like if you really forced it to, but it's very difficult for it. So, this would be like your your brain, your knowledge base, right? From which you can extract insights and chat with, you know, either all the sources or you can just select specific sources if you just wanted to dive into two or three specific papers and compare them, that that's also possible. Like, you could, like, you know, look at these three papers and just ask that, like, you know, for example, how do the results of these studies differ, right? And then you've got the answers in here, which is fantastic, right? And it can help you to do a lot of other things, right? It can, like, generate an an audio file as well, so that you can listen to it on the go. You can create videos, mind maps, slide decks, all that kind of stuff which is super useful. But for me, you know, this is really the brain. This is really the knowledge. And the problem with using AI tools such as Gemini, Claude, or ChatGPT is they don't have the context. They don't have the knowledge that a professor in your discipline would have. So we're solving that problem with Notebook LM because we're giving the knowledge to those tools, right? When you had to Gemini, you know, your general research assistant, I suppose, right? That can then use that knowledge that you plug in. And the way we plug that knowledge in is we go to the plus sign. And rather than upload files in here, we can just connect it to the specific notebook that we created. The advantage of that is that, you know, the context that it gets and the knowledge is much bigger. Because otherwise, there is a limit to how many files you can upload. Whereas with Notebook LM, we've created that knowledge base of, you know, up to 300 papers in one notebook. And you could, in theory, like upload several notebooks in here. And then this allows you to really use that knowledge in creative ways. So for example, we could ask Gemini to help you find research gaps based on the studies in Notebook. I mean, the possibilities are endless, but this is just one example, right? And then we, you know, it identified several research gaps. I mean, to be frank, of course, you know, this prompt isn't that good. The more detail you provided, for example, you know, are we sure that Gemini actually knows what's the process of finding the research gap? Well, maybe not. So we can give it more detail, right? And tell it who it should be acting as. But even with that, like notice what it found. And those gaps are, you know, a pretty good, actually. Like with me knowing the the field, they're actually pretty good. Like, it's not hallucinating. The reason it's not hallucinating is because we gave it the knowledge. Right? So, it's going to base it on that brain that we've created.
Are you ready to implement these strategies to publish research papers in high-impact journals in your discipline? Are you a professor, a researcher, or a PhD student who would really like to advance their career, make a really big contribution to the field by publishing more papers in better journals while actually working less and enjoying the whole process? Then I've got really good news for you. I've just opened some slots in my calendar, and you can book a free one-to-one consultation with me. We will dive deeper and identify the specific challenge and bottleneck that is blocking you from achieving your full potential, and then we'll also clarify your goals. And then at the end, I'll outline an action plan for you that will help you to achieve all your academic goals, publish more papers, [music] and advance your career. If this sounds like something that you want to do, book the free one-to-one consultation right now. The link is in the description of this video.
So, that's the second level. Like, it's like a general personal assistant. What we want to do then, level three, is to create custom AI assistants. In Gemini, they call them gems. So, Gemini is actually on the free plan, so you don't even have to be subscribed to the paid one, right? And the way you do it is you go to gems, right? And then you you create your gem in here. I I've already created like really an army of little AI minions, right? I can show you exactly how they work. So, for example, finding journals is a very repetitive task that you have to do over and over again every single academic year. And the instructions for that task are exactly the same. Just copy it into Word so that you can see it more easily here on the screen, right? But basically, you know, with a bot, we we give it a specific persona that we want it to to have, right? And then we give it an interaction protocol, so how is it going to interact with it? And then, you know, we've got an execution process. And basically, you know, these step-by-step instructions, right, allows to get really tailored input time and time again, okay? And you can create those bots for anything that you want, really. So, think of any task that you repeat several times in a semester as a researcher and create a bot for that. It will act as a specialized research assistant. And the cool thing is as well is that in here you can actually give it knowledge, right? So, maybe, you know, you follow a specific pattern when it comes to analyzing data. Well, if you're creating a bot to analyze your qualitative or quantitative data, whatever that might be, and there is a specific protocol that you know the bot needs to follow to get a good results, well, you can upload that protocol in here as a knowledge base. You just add it right in here, right? And this ensures that the bot, you know, actually follows that. Now, if you're not sure how to create this, a pro tip is to actually create a bot that is basically a prompt architect. So, I've actually created a bot for myself that helps me to create more bots, okay? Which is pretty neat because now whenever I like I have an idea or I, you know, I'm trying to respond to the reviewer's comments, it would be useful to have a bot for that, well, I go to the prompt architect and then it helps me to actually create those prompts and create those bots and then test them. So, that's level three.
So, so far we've got knowledge base, which is Notebook LM. Then we've got, you know, the general research assistant, which is Gemini or it could be ChatGPT or Claude or whatever. And then we've got the third level, which is the specialized research assistant, right, which are those bots, okay? Now, level four would be to use specific AI research tools. And I think currently probably one of the best ones, if not the best one in 2026 to use is SciSpace. Well, the reason why I'm recommending to use a specific AI research tool on top of the other three levels is that this gives you, you know, additional powers, if you will, because this has already been designed with, you know, specific AI research tasks in mind that a typical researcher needs to conduct. And it really unlocks that extra level that you probably won't be able to unlock even with those research bots that I've just shown you on level three of our system. And the the reason why I'm recommending SciSpace here, there's by the way a link if you want to sign up on SciSpace paid program, there's a link in the description. We also have a discount code. And but the reason I'm recommending it is is really to do with its agent gallery. I I just think that this unlocks so much for you, it's not even funny. Like they've got over 1,000 agents in different disciplines, right? So, if we go to biomedical, just look at all these different agents that they've created, right? If you go to chemistry, you've got all these agents in here as well. And so, they have been specified to that particular discipline. So, you know, I've shown you how to create an AI bot for finding journals, right? That was an example I gave you. They've got one here specific to chemistry. Okay? They also have, you know, specific bots that allow you to conduct chemical analysis, for example, which I think really is is fantastic. It saves you so much time that it's not even funny. And it's already been pre-built for you. So, it just unlocks that extra layer of more agentic AI use, where it's able to connect different applications together and run more complex tasks such as, you know, analyzing complex data, preparing graphs and figures, you know, doing a whole systematic literature review. If we go here, there's an agent that can do a whole systematic review from start to finish, right? Or it can map the evidence, right? Conduct an advanced Google Scholar search, you know, prepare a Prisma flow diagram, all these things. So, basically, you know, it it acts as that next fourth level of your AI research system.
Now, if you've watched this video and you're like, "Well, this this is really cool. I want to build such AI research systems for myself. I want to be publishing three or more papers every single year." Then in the description, there's there's a link somewhere there where you can watch a free video case study where I show you exactly I break the whole system down of how me and our clients are able to publish three, even five Q1 papers every single year. And you'll be able to book a free one-to-one consultation as well. We'll dive deeper into your specific challenges, see how we might be able to help you achieve those goals. Now, of course, AI on its own is not really going to give you Q1 papers. You still need to understand how Q1 papers are actually put together. So, that's why in this next video, I show you exactly what it takes to publish in Q1 journals. So, watch that next.