Speaker 1: There are so many different types of AI research assistance, they come in loads of flavours, but I think when I'm doing research, there's three areas that I really like to look at, and this video will go through the best AI research assistance in those different areas. The first one, dare I say it, makes research very, very fun, and that is when you're using semantic search to find particular things. Now, there are different types of services. These are my favourite. The first one is consensus. Consensus is really simple, because you ask it a research question, and it will give you a nice summary, and also a list of papers that help you understand that research question. So here, we'll just use one of their example ones down here. We want to synthesise, and we want co-pilot, and let's say we want do tattoos impact behaviour? Oh, do they impact behaviour? So as you can see, you get a summary here, which is just a little summary AI generated of the general consensus of the papers at the time. You get a consensus meter, and you get all of this. You get a co-pilot that gives you an answer, and it's all referenced really great. Another tool you can use is illicit. The same thing applies here, where you just go in, and you can find papers, academic peer-reviewed papers that answer a research question. So I can go in and ask a research question such as, and then illicit will go away and ask the literature that question, see if it can come up with an answer. It's a pretty weird question, so let's see what it says. Here, it's just got, research suggests that the smell of feet, particularly when associated with stress-induced chemo, can be aversive to animals. There we are. Even the most weird questions, you do get a response, and as you scroll down, you've got the paper. You can also do filtering in various ways, and so I really like to use this as a first port of call when I've got any weird questions, and I want to know the scientific answer to a question. The last one is SciSpace. SciSpace is a brilliant AI research assistant that you should consider using. And by the way, don't worry, this is getting pretty confusing with all these tools. Go check out my course, where it's a really easy way to create an AI toolkit specific for your research that can help speed up everything you want to do in your research or PhD. So go check it out. But here is SciSpace. SciSpace is a fantastic tool because it is an AI research assistant that does almost everything you would want it to do. So here I want to go in and say, are beards better for health? Let's have a look at that, and then it will go away and start to look at the literature. So here we go. It's found five papers. It's going to give me an insight, so there's a summary, and then we've got 10 papers down here. We've also got columns that are AI generated that give you a little bit of a summary of particular papers in each row. So these three tools are really fantastic if you've got that little idea, and you're like, I want to ask the research. I want to ask the peer-reviewed research this question. But we don't always want to start with the peer-reviewed papers. We want to go elsewhere, and there's one tool that I love. Check this one out. If I want an AI research assistant that starts really broadly, and I don't really know where to start looking, the first place I always go to is Perplexity. I absolutely love it. Essentially down here, you can go here and ask any question, and then it will go out to the web, and look, here is the focus tool that I really like. We've got all academic writing, Wolfram Alpha, YouTube, and Reddit. So it will go to these particular places to find an answer to your research question. So if I'm unsure about where I should start, maybe it's not a super academic question, but rather it's something more general where I do want a website to answer it. I will come to Perplexity, and I will just type in my question. So the first thing it does in Pro is it tries to work out what your question really means, which I really like. So what type of relaxation are you looking for? I want to say physical, and I'll send that. Then this Pro search will go out and essentially look all over the web for an answer to that question, and it will give you references as well. So you can see here it's got sources. It's got North Shore. It's actually got this, which is a peer-reviewed paper, which is great. So it is mixing it all in. It has got this website, and it's got 17 more sources, which is just so much. We've also got the ability to search videos. It's got all these images here. So overall, oh, ad block. I don't want you. But overall, all of this is just a really great way to start if you need to start somewhere with your research, a great AI research assistant that isn't solely based on the peer-reviewed science. Love it. Another way that you can start your research with AI is if you already have a collection of peer-reviewed papers that you want to explore further. This is the third kind of flavor of AI research assistants that I really like, and check out this, Enargo Read. It's a relatively new AI tool, but essentially what it does once you've sort of like logged in is you can add your literature, and you copy and paste all of your PDFs here. You can drop them there, which is great, and that's what I've done with some of my peer-reviewed papers that I've produced throughout my PhD and my postdocs. Now, here's the thing is that we want to now explore a little bit further. We want to use this assistant to go and explore what we've already got and go beyond where we're already searching. So this is how we do it. It's not super intuitive on this platform at the moment, but essentially we've got these options. So let's say this paper, multi-layered approach to polyfluorine water-based organic photovoltaics. So down here we've got summary, references, and related papers. So this is where we can start exploring and using this AI research assistant to broaden our horizons. There we are, that's what I want to say. So here, summary. You get this pop-up, and then you get all of this summary. So this is a generated summary of each section of the paper. Now, one thing you can see is I can highlight it here, which is great, and then we've got chat with co-pilot, pin this statement, related papers, view institute resources, and just so much more that we may want to use, which is great. So introduction, experimental. So here I can go in and get a quick snapshot of each individual section of the paper, which I really like. We got key insights from each thing. So this is the research goal, the research context, the approach. So it's a nice sort of place to start exploring the papers you've already got. Then we've got figures, we've got tables, and we've also got info, which is all of the rubbish stuff that goes with it. But this is where it gets even more powerful, in my opinion, as an AI research assistant, is we go to open literature, we click up here, and the great thing is that then we get into this view on the website that allows you to go deeper into each individual paper. So this is the PDF of the paper that's just popped up. We've seen the section-wide summary, we've seen the key insights, but here it is, critique. So here now it will provide you with critiques and arguments. What are the premises and assumptions? What is the context and background? And it will really help you kind of understand and create a little cheat sheet of your own thoughts about particular papers, which I really like. So here we are, this is the literature, and then over here in Copilot, this is where you can start chatting with this paper. So you can select text to get an explanation like before, or we can summarize

Speaker 2: the paper, or we can ask it any question, really. Andy Stapleton, Andy Stapleton, Andy Stapleton. Always thinking, which means that it's almost certainly me.

Speaker 1: The tension's killing me. I'm sorry, can you provide more context to clarify your question? Oh, be quiet. Anyway, you can ask it proper questions. So now I'm back on where I put up all of my references, and here is where I can now push the boundaries of what I know and what I can find. So here I can go related papers. Each paper has got this one underneath, so you click related papers. Then you can sort by related research, who it's cited by, survey papers, or review articles. Quite often I like review articles because it gives you the chance to see what a current state of a field is like, rather than delving into individual papers. That can get a little bit confusing and muddy the water. So early on, review papers or review articles, and now we'll go away and find related review articles. I've used this a few times, and I think it's really quite powerful, and it does give you a way to explore a little bit deeper. So here we are. These are all of the different papers it recommends, and you can add it to your literature. You can attach it to a particular paper if it's something you want to attach to something that's already in your literature list. But overall, this is not such a bad way to start exploring and reviewing the papers you've already got and expanding the amount of literature that you've got for your project. Great. The last thing you should do, and the last type of expanding your horizons type of AI research assistant, is one where they map research. Here I've got connected papers. I like connected papers because it's free, and I use this like this. When I have a seed paper that I just want to know the related, the derivative works, what happened after this paper, particularly important if it's a really old paper, I put it in up here just using this bar up here. Then it creates a map, and I always go to derivative works. This is pretty much the only way that I use this tool, by the way. I just click here, derivative works, and then I can see who did what, when. So I can see that this is a relatively recent paper from 2023, and then we've got 2020, and so then I can go on and click. I can go and find these, open these papers, and start reading them. So that's a great research assistant for you if you already are in a particular research field. You just want to look at like the derivative works of a particular scientist, researcher, or a particular project that you know will be suitable for your research field to know about. That didn't make any sense. You know what I mean, don't you? Alright, if you're a research assistant nerd like me, go check out this paper where I talk about how to use consensus AI in detail and make it super powerful for your research.

Speaker 2: Thank you.