Speaker 1: A new era of AI for research has begun. That is, in the past, we've had to use multiple tools in academia, and now we're starting to see tools where everything is combined into one, and that is exciting. And the one I wanna talk about today is Petal, or P-et-al, et-al is what you put at the end of like a list of authors to say and others. This is generative AI to chat with your documents, but it does so much more. If we go down to what it can do for academia, I think this is where you get a glimpse of what it can really sort of like do to power up your research. First of all, we've got very simple, simplifying a literature review. So you can put in PDFs, you can put in references from a range of different sources, and we'll talk about that in a minute. You can actually store and organize your research library, and you can collaborate with others. So I think this page actually doesn't do it justice as to what it can actually do for your research. So let's jump in. I've actually created an account, and in here, when you sign up, I'm on the free plan at the moment, and this is really only to test it out, but the pricing is really, really sort of great for academics. If I go to pricing, you can see here that it's only $2.55 per month with an EDU account. So if you're on a student account, you can get this for almost the price of half a coffee. Once you sign in, you have got this dashboard, and this is where the power is. You can upload a load of different documents from wherever you want. I uploaded it from my computer, but you can also import via Bibtex, which is brilliant. You can also import via DOI or whatever sort of identifier. That is really cool, especially if you're sort of like out on the web and you're just sort of like collecting for that initial literature review. And coming soon is import from Cloud Drive. You can add an entry manually, and you've got Web Importer. So all of these are really great ways, and it's exactly kind of like where you find research. So you can get it all in one place, great. Once you've uploaded your documents, you go to this Documents tab, and you can see I've put in all of my, well, some of my papers and my PhD thesis. And so I wanted to see what it could do with this massive file, and as well as all of my other papers over my PhD and postdoc years. Then you've got two other tabs, AI Table, where you create a table, and that is really powerful. I'll show you what that looks like in a minute. And then you've got Multi-Doc Chat, where you can create a new chat and speak to specific sources that you've uploaded. So it's not just like saying, here's all of my sources, I want to talk to all of them. If you've got different segments to your PhD or your research, it's a great way to say, I'm now chatting with this bulk, I'm now chatting with this sort of data set. All of that is super powerful. So let's go through. The one thing I like about this is you can actually talk to individual documents as well. So if I want to talk to this one, I click on it, and I get this panel on the side. It tells me properties, and then essentially all I have to say is Ask AI. This opens up another panel, and then you've got all of that power from other tools in one place. So here I can annotate, I can put comments, and then I can talk to it using AI. So let's have a look. I just want to annotate this bit. Oh, actually, I am on, this one I'm currently on is a highlight rectangle, but let's say I want to select text. I can go to create an annotation, and I can color code it. So I'm like, read this. There we are, that's great. So I've got to read this, and then I've got different highlight rectangles. So for example, if I've got this schematic, I can go in, create a little highlight, create a thing, and say, do this in the lab. Great, brilliant, it's all there. And then on the side here, we've got chat with that document. So I want to chat with this document and say, what are the limitations? Oh, got to spell that right, limitations. Come on, Stapes, you can do better than this. All right then, then we're going to ask the AI to summarize the limitations. What are the limitation? Brilliant, well done, me. The provided document does not mention any specific limitations. Oh, great. So apparently it doesn't kind of, it's not able to infer. All right, so I'm not completely happy with that response. Let's see if I can say, okay, what are the key points from this study? Come on, you've got to be able to do that one, go. Okay, the key points from this study are these things. Okay, good. These notes are based on the provided document episodes and not encompass the finding of the full study. Interesting. So if you highlight stuff, apparently it's only asking the bits that you highlight. Hmm, didn't know that before starting this video. You can also, you've also got an outline tab over here. So overall, I think this is a good way of, yeah, just sort of like delving deeper into individual research articles. Let's go back and have a look. So here are all of the articles that I've uploaded and I've got my AI table. This is interesting to me because it sort of like emulates something like illicit.org where you create a table, you can manage your sources, but here I've got all of my stuff in there and then I can create a table where I say, okay, here I've put one before and it's a summary. So here I've said it should be a summary and the AI question that I want to put in the column is what are the main conclusions from this document? And then it appears and I can go, okay, get answer on this one. I'd like a button that just was able to kind of like go, you know, do all of it, but at the moment it's just individual papers that you upload that you have to go in and put, get answer. Overall, a little bit annoying, but I think it's a step in the right direction. Maybe if you upgrade, you get a button that just says do all of them, but you could easily go through your AI credits, which is probably why they're not giving you that opportunity because you could just burn through them so quickly. So let's have a look. So in this table, I want to add a question. I want it to be UK English. I want this to be, I want it to be, let's see if we can do limitation. No, even better. I'm a genius. Let's do further work. That way they can kind of like help you decide what you're going to do with your work if you've got someone else's research you're looking at. Okay, what experiments could expand on this work? That's interesting to me. So I'm going to go in here. The one thing that I found really funny was the summary for my thesis, it just said, I don't know. Yes. I guess it's because the thesis is like 255 pages, but I love it that it's just like, eh, I don't know, you deal with it. Okay, good, good. Maybe if you pay for it, it gets better. I don't know, but that was funny to me. Here we are. So some possible experiments could expand on this work involved, and it gives me four options. That's pretty good, especially if you're trying to come up with a research proposal, you're trying to find a gap in the literature, and the one thing I love is that you can put any question you want, any AI question into a column, and you can make sure that you answer it for all of these different studies you've uploaded. Brilliant. Now, if you don't want to go through manually, just click and sort of like come up with your own questions. The multi-doc chat, I think, is where it's going to become really powerful, particularly if you've got that large body of work you've been sort of like scooping together from different sources. You can put them all up here. So here you can click and say, I only want to include these sources, I'm just going to leave it with all of them, and you can see here I've asked it, what are the best directions to take solar cell research based on the documents provided? And it's given me all of these options. It's given me five options based on the individual articles that I've asked it to sort of like read over. So I really like that. Now let's have a look and just say, what are the best solar cells in these studies? And now we wait. All right, so it says it's not explicitly the best solar cells. So the AI at the moment is a little bit lacking in terms of its interpretation of the questions. So here it's saying like, okay, these are the different types of solar cells were made, but it hasn't extracted the efficiencies. So there's still some room for improvement. I think this is what happens when you try to combine a load of different tools into one places, you end up doing a load of them like okay. But I think Petal or P-E-T-A-L are on the right track. Watch this space because I feel like they're going to get better in this area. Let's ask it a follow up question and say, what are the efficiencies of the solar cells? Okay, okay, okay, I've pushed it a little bit and we actually got numbers, which is good. So here we can see the multi-layered approach, the power conversion efficiency ranges from 0.0039 to 3.9. So we can get the information out there, we just need to talk to it a little bit more, be really specific with what we want. I like that, okay. It's better than I thought it was a moment ago. And here it's pulled out this, 0.4 for devices with an aluminum cathode and 0.8 for a calcium aluminum cathode, brilliant. Okay, good, it's pulled out information. That's the sort of power it needs to have if it's going to be useful when it's talking to multi-documents, especially multi-documents that are full of really rich data. So this is good, okay. We can relax, it's better again. So there we have it, that is how research is changing in the AI space. We are starting to see tools that stop sort of like just doing individual things and are trying to put them all together in one package. I think we're only a couple of years away from having an AI tool that is sort of like a research assistant helps you across a range of different fields and that could really sort of like increase your productivity because it acts as a single point research assistant for you. Let me know in the comments if there's any out there that you've tried, I'd love to hear about them. And also, if you like this video, go check out this one where I talk about the AI tools that are so good for academia that it feels like cheating. It's a good watch. So there we have it, that's everything you need to know about P et al or Petal. Let me know in the comments what you think and also remember there are more ways you can engage with me. The first way is to 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, my TEDx talks and more. It's exclusive content available for free so go sign up now and also remember to go check out academiainsider.com. That's my project where I've got blogs, I've got resource packs, I've got e-books, I've got forums and courses are coming soon and everything is over there to make sure that academia works for you and not just your supervisor. All right then, I'll see you in the next video.