Speaker 1: So can AI do research for you now? In the past, ChatGPT has really struggled because it doesn't have internet access. They gave it to us for a little bit and then they took it away. How dare they? Now, there's a load of other services that have come out that could potentially fill that gap and they are promising academic research. Let's have a look what they mean. So this is what GPT-4 is doing at the moment on ChatGPT. I have essentially just said, find the current state of organic photovoltaic devices in the academic literature. Now, I know full well it doesn't have access to the internet. It used to and it used to also provide links. Let's see what it gave us. So it just said, as of my knowledge, cut off. Yes, yes, yes, we know. Here are the advancements and key trends. Okay, great. It's just sort of like generic stuff. That's not really useful for anyone who's doing academic research. And then I came across Silatus or Silitas. I don't understand why it's so ambiguous, the name, but it says here, let AI conduct your research. And I was like, yeah, whatever. It's just going to be like normal boring research. But no, it says academic research. Really, Silatus or Silitas? Can you do academic research like me, a PhD trained scientist? Let's have a look. You can see here that you end up with kind of this interface once you log in. I'm on a free trial at the moment, but essentially there's these options. You get general and academic. Academic, yes. And here we've got fast and precise as well. Now, you type in your thing there, you push go, and it takes about sort of like six to 10 minutes as far as I could tell. Now, I've done this previously, and you can see here I typed in exactly the same prompt, which was, what is the current state of organic photovoltaic devices? And I wanted to go away and have a look at the field. So it had a look at all of these different sources. This one, this one, this one. We'll take a little bit of a look to see what age the papers are. This one's 2019, so yeah, whatever. 2019, is that the latest? Ooh, 14.7% efficiency. Let's have a little bit of a look. 2019, okay, so overall it's not really giving us like the amazingly up-to-date stuff, but is there a better one? Stay around, because I think there is. Let's carry on to see what happens. So after giving me all of these pub.acs.org, so that's where they all are. I mean, this one's RSC, ASACS, and then it gives us this kind of like summary document at the end, which I kind of like. So let's open this up and have a look. It says, your input is, what is the current state of organic photovoltaic devices in the academic literature? That bit's very important. So here it gives us the factors, the strategies for improving recent developments, and then it gives us this, a simple rundown of each individual source that it finds. So here it's got context, context, context, context, and then the sources at the bottom. So is this academic level research? I don't think so. Overall, Sellitus, yeah, you've done an okay job. I wouldn't pay for this for academic research though, even though it says it can do academic research. There are better ones. Let's check out those. The next one you need to know about is IOMNI, IOMNI. Why, why are the names so difficult? And here you can see you've got your business intelligent sidekick, but essentially you'd end up with autonomous market research, and that's essentially what we're doing. So I can click on this research button at the top here, and you can see I've asked yesterday, what is the current state of organic photovoltaic devices in the academic literature? Let's click on that to see what it came up with. Now, the one thing I really like about this is that it really understood the brief. If I click on these, we can see that this is a 2022 paper, much better than the thing that said it would do academic research. But essentially you can get a look to see its internal monologue, and I like the fact you can see the wheels turning as it's looking at things. So here it says I'm looking for the latest advancements, then I'm looking for the current challenges, then I'm looking at the efficiency rates. So it's decided that it's got all of these subtasks that it needs to do. Now, do you remember agents? Do you remember AI agents, these sort of like little bots that go into the world and find information for you? This is what it is. It has improved significantly and dramatically over the past sort of like few months. And here we can see that it's looked at materials as well, it's looked at potential applications, and it's given us, in my opinion, a much better summary of the current literature than salacious or silatus or however you say that, I've kind of forgotten. But here we've got key takeaways, and we've got the introduction, we've got efficiency rates, which is up to date. We've got materials and device architecture, stability and lifetime, manufacturing, commercialization, conclusions. And the one thing I like about this, as far as I could tell, it was free. I didn't pay for anything. Staying in the world of AI agents, i.e. little bots that go out into the world to find out that information you're searching for, there is this service called Cognosys, much easier to say, I like that one. But essentially once you're logged in, there's a free version as well where you get a certain number of bonus credits. You can actually then go to see what agents you've looked at in the past. And yesterday I did exactly the same prompt to Cognosys. So here I can click and we can see that it was using the GPT 3.5 model. So if you pay for this, I reckon it could get even better. But essentially it's gone through the literature and the one thing I love, it doesn't just stick into one ecosystem. It's actually found papers from a multitude of different sources. And it's giving me a rundown of what it's found, a rundown of what it's found here. It's used two articles. But the one thing that I really like about it, if you click on here, it opens up where you get the information. So here it's a 2023 paper. There we go. Oh my God, 2023. And it says advances in organic photovoltaic cells, a comprehensive review. So it's giving me exactly what I wanted, which was an up-to-date review of the current state of organic photovoltaic devices. So I think that Cognosys, even though it's delivered only two papers, it's actually sort of like fulfilled the requirement of the prompt, which is the latest advancements. And I'm not saying that this will replace a literature review search or that it will sort of like be able to produce stuff you'll be able to copy and paste into your thesis or wherever. It is not going to be able to do that. And to be honest with you, I have seen the AI agent world come into its own over the last few months. And I'm pretty excited of how well it is performing now and where it's going to go in the future. So check out these agents and these research tools for your research field. And let me know in the comments if they work for you. It's free to give it a go. You get a free trial with some of them and it could really save you hours and hours of work. People get a little bit annoyed when I talk about AI tools, but essentially here, we're not saying that these AI tools do everything for you and that you can just copy and paste into your paper thesis or wherever. You still have to do the sort of like critical thinking because the AI cannot do that. It does not have an intuition that you build up as a scientist and a researcher. For example, when you read a paper, once you've built up that intuition of what a field is like, what's the sort of like leading edge of ideas and discovery, you start to have this kind of like bullshit sort of like sensor and you can be, okay, like this is not really fitting in with what I'm reading at the moment. Or you can say, okay, this is a bit of a rubbish paper. It doesn't really give me anything new. See you later. That still exists. That still has to happen. And it should happen for everything you read and you want to cite in your thesis, your paper or wherever you're writing. But it does not mean that you cannot use awesome tools that are coming out and getting better by the day and it's important you stay up to date. So just subscribe to this channel and I'll do my best to keep you up to date of the best tools for certain activities in the academic realm. Easy. That's a good deal, isn't it? So there we have it. There's everything you need to know about the current state of AI tools for doing academic research. Remember, it's not research. Really, it's just mining the internet and the world for different sources related to your research. You still need the critical thinking. You still need to sort of build up that intuition around your field. But this makes it a little bit easier to put those feelers out and grab some stuff. And it's a great first touch point for a lot of researchers if they're entering a new field or they just need to grab that little bit of information from somewhere. So let me know in the comments what you think. And remember, there are more ways that you can engage with me. First of all, you should sign up to my newsletter. Head over to andrewstapeton.com.au forward slash newsletter. The link is in the description. And when you sign up, you'll get five emails over about two weeks. Everything from the tools I've used, the podcast I've been on, how to write the perfect abstract, and more. It's exclusive content available for free. So go sign up now. And also, go check out academiainsider.com. That's my new project where I've got my eBooks, my resource packs, the blog, the forum. Everything's over there to make sure that academia and your PhD works for you. All right then, I'll see you in the next video.