Five Free AI Tools Researchers Should Use in 2026 (Full Transcript)

[00:00:00] Speaker 1: Hey, Professor Stuckler here, I'm going to do something I normally don't do. I'm going to share my top five AI tools list for the year 2026. Now, many of you who follow the channel know that I spend a lot of time to spelling out AI failure modes and unethical ways to use AI so you don't get in hot water, especially because I see a lot of students going down rabbit holes trying to use AI to accelerate their research when they haven't got the fundamentals right. And that just leads to a big mess. Garbage in leads to faster and bigger garbage out. And the five tools I'm going to share with you today, I don't want to say they're indispensable, but I use them all the time. And they're the ones we use inside our FastTrack mentorship program. They're not what you think. The number one tool is not ChatGPT. So I would encourage you to stick around because some of these you haven't seen. The other great thing is they're all free. I don't get any sponsorship or benefit from any of the tools that I'm recommending. Just take that with a pinch of salt when you're looking at some of the other so-called influencers out there, because often behind the scenes, they are profiting on what they recommend to you. So with that, I'm going to show you these tools. I'm also going to show you my favorite use cases for them that you can implement today on your projects. So let's dive in. We are going to go reverse order number five on our list. And that is coming from Google's Gemini, the fun word to say, NanoBanana. And NanoBanana is an image editor or inventor. What's great about it is it really can produce high-quality publication-style figures. I've even taken a notebook before. I have my notebook here outlined for the video. Drawn up a figure or a diagram and loaded that up to Gemini, and it makes a very nice figure. The other thing that I've done too is there might be an existing figure that I need to redraw. I can pop that into Gemini in a prompt. I've actually got an example right up here on the screen where I wanted to highlight some regions neurologically that were associated with autism. And there's still some things I have to clean up. I might have to take it over to Photoshop or something accessible like Paint or even Canva to clean it up. But it can help me produce, edit, or even produce very quickly publication-style figures. The one thing you have to be careful with this app is the reproducibility. So an editor or a reviewer could ask for changes down the road, and it can be difficult to make a perfect copy of your image. And if you upload the same image that you have and try to make tweaks to it, sometimes it still isn't perfect with manipulating text. I find that's better to edit sometimes elsewhere on another app, and multiple prompts on the same image can start to degrade the quality over time. But overall, it's pretty good, and that's why it makes my top five list. Let's go next to a tool that is actually part of our tech triad that we recommend to graduate students and researchers across the piece, and that is Grammarly. I mean, it's just indispensable for grammar. If you're working with multiple co-authors, you're going to have lots of track changes in your documents. Things are just going to slip in, and you need to make sure you don't have any rookie errors in your manuscript. When you submit it, those rookie errors, I mean howling grammatical errors. It just gives this impression that you did a cursory job, when you probably didn't. But I do see researchers kind of exhausted, gasping as they cross the finish line, and they end up with a spelling error in their title, or the first sentence is off, and you'll get desk rejected, or your supervisory just will have a quick gestalt to think this is probably rubbish, and not take it as seriously as they should. Now, Grammarly's been amped up with AI features, kind of like some of the other tools out there that I don't recommend, that I'm not going to name, but it can work in-house and live inside your Word document. And so if you're paraphrasing, and you're going to paraphrase in the right way, check out this video here for how to paraphrase in a good way that's ethical. Grammarly captures one of our two paraphrase methods that we advocate very nicely. Just always, of course, if you're using these AI features here, check it for scientific accuracy. But you don't need the AI. I do recommend Grammarly to everyone, but it can be useful here, and it is my recommended tool for paraphrasing. I used to recommend ChatGPT for it, but the fact that it now lives nicely inside your Word document, you're using it anyway, makes it number four on my list. Let's go into a next component, and it is Google Scholar Labs. So Google Scholar is something that I recommend for everyone when doing their searches. It is just bread and butter. It is classic. It's tried. It's tested. It has great coverage. It's not just limited to open access coverage, which is what a lot of the AI search engines are fishing from. So I highly recommend it. Its search algorithm just works really well. It will help you forecast the impacts of your papers. It'll help you map out your topic spaces. Just a really good tool. And so I just did a random search to do a comparison. What the AI amped up version Google Scholar Labs is doing, it's free, is it helps you, again, it has a slightly different algorithm to optimize for what might be a research question. Now, I haven't fully validated this comparison. I've got one on the screen here so you can compare. Here's a classic Google Scholar search on social capital diabetes. And you can see I'm getting a relevant search. This will be relevant to me. So your search here will be different from mine. It usually is optimizing in my experience for recent papers, but also classics, so high citations, things that seem foundational in the field. But I actually find when I look at the output from Google Scholar Labs that the results it found feel more relevant. And it gives me some justification and explanation for why it's more relevant. And so I found this for just exploring the literature to be slightly better than Google Scholar. By no means is Google Scholar without AI perfect. But I find using these in combination is a really helpful way to map your topic space in-house, one free platform using AI and without AI that's drawing on the same broad data sets and data points of the literature that it has access to. And that's why Google Scholar makes number three on my list for 2026. Now, as a side note, one thing I do commonly see with researchers who come to me who've been struggling is they fall into this trap of just thinking tools are going to solve them being stuck. And usually there's a deeper problem going on. So when I see someone who has a tech stack of, I mean, frankly, more than five tools that they're using to try to solve things, they're often creating more frictions. And they're often using these tools to hide from the fundamentals. They're looking for a quick tech magic bullet fix rather than doing the hard yards to really learn. I just find it's faster in most cases in workflows to not use AI than to use AI. There's a fascinating study that came out that actually looked at developers who used AI. And while the developers thought that they were faster with AI, they're actually spending more time cleaning up the errors that AI was introducing into their code. And this is really what I see on the surface of AI use. It changes the surface of where errors occur in the research process to where it helps people get past early frictions that would normally catch errors in the old pre-AI world. And they now start manifesting deeper later, more structurally when you've lost a lot of time. So again, I'm just pausing here in the middle of this video to do a sense check with you that you might want to have a look in the mirror if you are building up a massive heavy tech stack and trying to hide from some of the deep work you need to do with magic bullet quick fix AI tools. All right, let's go on. Number two. And of course, this is going to have to be on any list. Any list is incomplete without it. You just can't ignore it. And that is ChetGBT. Look, it's foundational. It's for many of you where your AI experience started several years ago at the time of making this video. And it works. It has a lot of use cases. I want to share with you a couple upgrades that I use in ChetGBT. One is I recommend that you go into projects. Create a project. So you're not just using an individual prompt, but create a project for each of your research papers. That's going to have better coherence and memory across your chats. Again, it's constantly evolving, but that works really well. I've also used the pro model. Unless you're doing very highly specialized things, you don't need a subscription to pro. And the free versions work well. If you do get the light version at the time of this video, current model of ChetGBT works just fine for most common purposes. And the reason is for the best purpose I like is to simulate peer review. And this is my favorite use case for ChetGBT right now because there's been an analysis recently that found AI content in 53% of peer reviews. That is remarkable and it's perhaps not surprising that our whole edifice is built on free work in academics. And so people are doing AI peer reviews and then humanizing it later. Deeply concerning. But what that means for you is that they're probably going to use ChetGBT because it's the most common one right now. And so you want to simulate that and you want that to be free of context. So don't put that in your actual project when you're simulating the peer review. Maybe even use an incognito browser in Chrome, for example. And use the prompt. I'm sharing some prompts below that you can lift and use. And this is really in our principle to not have AI be a false friend but force it to challenge you to give you critical feedback. And often this does spot some things that will come up in peer review. And you can just nip in the bud. It's bulletproofing making your paper stronger as you go forward. Again, interacting with ChetGBT. I used to recommend certain AI apps in the past like Syspace, which is still a great app at the time. It was one of the best where you could upload PDFs and chat with PDFs. ChetGBT can now do that inside projects. It's just not as good as some of the other ones available. So this is, again, taking together why ChetGBT is not number one on my list. It has some fantastic use cases. And I still recommend it. I still use it almost daily. But it's not number one. And so last but not least, recently, you know, if I did this last year or two years ago, it would have been ChetGBT number one. Now it is Google Notebook LM. I remember my video editor, he pushes me a lot. Lev, I know you're watching this too. I keep saying, I gotta look at Google, gotta look at Google. And it just wasn't there. I think it's crossed a critical threshold and now has a lot to offer. And my research projects now live here. So I've got an example here of a project where I've loaded up all my research. Sources and my library. It's a project that I'm doing on causes of autism. Really fun project. And what I love about this is I can really poke and prod not just one paper, but all my PDFs. So I can have a conversation with my specific research projects body of literature. You can do cool things here from making mind maps, which other AI tools try to do is all built in-house. You can try to do video overviews, infographics. I don't think you really need that. Sometimes it can give you some interesting ideas for figures. It can perturb your logic and poke and prod your data. Sometimes something researchers will do is try to look at things from different angles and through different lenses to get insight. And I find Notebook LLM is a great tool for doing that. There's a great feature here that you can make a customized podcast for your research and means that you can activate some of these dead spaces in your day to learn and really get familiar with your literature as you go. So doesn't really work as a writing tool, but I don't want you to do that anyway with your writing. And it's okay, but I don't want you to take a heavy footprint beyond really grammatical edits. You want to work with a writing system like our peer writing system. Video link here. I often say, give me five minutes. I'm going to transform your academic writing for forever. It's one of those things that once you see it, you can't unsee it. And it will change the way you see your writing today. Check it out if you haven't already. But this is why on the whole Notebook LLM is just the top king of the castle right now. And it's free. It's all this power for free. I think they're trying to win the market with the chat GPT. I worry they try to commercialize this later, but tap into this now. It is the best place to have your entire project live. Guys, hope this was helpful. But I really have a sense. Check with yourself if you're paying for subscriptions and you're using AI tools to hide from some of the deeper work. I don't think at this moment. Convince me if I'm wrong in the chat below. You don't need anything beyond these five. Just don't. And with that, I know that's not going to earn me a lot of fans or friends. But I have no AI tool to sell you. I do have a take it back. I do have a fast track GPT that I'm constantly trying to refine. That's trained on our data. I'll leave that link below as well. You can check it out, which takes AI and overcome some of the main failure modes and turns your prompts into learning opportunities around research papers. So check it out. That will definitely help you, but it's 100% free. And guys, I look forward to seeing you in the next video. If you do want to learn more about AI failure modes, check this one out and make sure this is if you are using AI, doing it responsibly and not setting yourself up for these failure modes. See you guys in the next video.