How To Do A Literature Review (STRESS-FREE!)

There are so many awesome AI tools, but how do you use them together to write a literature review? Well, I'm about to

literature review? Well, I'm about to show you the ultimate workflow and this is what it looks like. So, we'll head over here and this is what I am doing in my mind every time I'm doing a

literature review. First of all, finding

literature review. First of all, finding literature, go and search, go and absorb that information. Then, I'm filtering it

that information. Then, I'm filtering it using AI tools. And by the way, there's AI tools in all of these, spoiler alert.

And then we do a full read of only some of the papers. The great thing about AI tools is that you don't need to read everything and then at the end we come out with an awesome literature review that you are proud of. So let's get into

each individual step and the sample AI tools that you could use for each one.

So the first thing you need to do is go and find literature. There are loads of ways to find literature. Ultimately

first of all though you need a reference manager. Zotterero is one of the best

manager. Zotterero is one of the best ones that integrates with a load of different AI tools, but you can use others if you want. But if you want to connect it directly, Zotterero seems to

be your best bet at the moment for a load of AI tools. And then we go searching, illicit, consensus, lit maps, and look at this good old-fashioned

Google Scholar. That's one thing people

Google Scholar. That's one thing people forget when we say use AI tools for academia research. We don't have to

academia research. We don't have to ignore all the old ways of doing it. In

fact, I quite like going to Google Scholar sometimes and typing in those key words to find papers. Then we've got connected papers. We got sci space and

connected papers. We got sci space and research rabbit. All of these places are

research rabbit. All of these places are where you can go to find research and just collect it indiscriminately.

Get all of that delicious stuff into your reference manager because the next step is really where the true power of the AI tools come into play. And you'll

see what I mean. So after you've got all of that delicious research into your reference manager, we need to filter it.

Now, we're filtering it for really sanity reasons because we cannot literally read every single paper that's ever been published in a particular field. So, we want to bubble up the

field. So, we want to bubble up the really awesome research that we can use to really sort of ground our understanding of a particular research field. So, now that's easier than ever.

field. So, now that's easier than ever.

In the past, you would read all of the abstracts, but now we can sort of shortcut some of that heavy lifting by using AI tools. So, filter the literature. We can use notebook LM size

literature. We can use notebook LM size and consensus. So we can head over to

and consensus. So we can head over to something like size where I've got a particular search query that I've used and this is everything it's kicked up now by relevance it sorts it. So we can

see here this is 92 out of 100. So

clearly I maybe want to read this one.

If I want to know something specific look I can look at the abstract and I can also add columns to look at all of the insights. So maybe I do want

the insights. So maybe I do want insights and I'll put that in and then it will go away create a column with all of these papers with insights. It found

so many many papers. You can see there's so many I can choose from. Ultimately

it's about filtering these down. Now I

don't want to read a paper that has got you know a really low relevance score.

So I'm making sure I read the top five papers up here just to give me a really good grounding. I can also use something

good grounding. I can also use something like consensus. Consensus is here and

like consensus. Consensus is here and you can see here up here key papers top contributors. I want to make sure I read

contributors. I want to make sure I read these top five papers right here. And

then it gives me all of the claims and evidence down here. This is making sure that you understand what the key papers are in your area. And the last thing you can use is notebook LM. You can put all of your references here. And then down

here, you can say something like, "Oh, let me know of the key uh findings in these papers and the most important papers from a particular all of that stuff." And it will highlight all of the

stuff." And it will highlight all of the stuff you need to know about the uh literatures. So those are the tools that

literatures. So those are the tools that I would use to bubble up the really important stuff for your literature review. The next thing is we need to

review. The next thing is we need to work out from the ones that we've filtered what actually they say. Oh, and

this is where we go old school. There's

no cheating yourself in this area, right? We need to make sure as academics

right? We need to make sure as academics we're holding on to the actual skills that are valuable for academics and researchers. And one of them is to read

researchers. And one of them is to read a full paper and get the right information into our brain. So we can start building that sixth sense, that

understanding of a research field that only comes from reading the research.

But now we can take all of the filtered researcher and we're looking to read between about 20 and 25% of the actual papers we've bubbled up. That is going

to give us a great uh starting point for understanding the literature. We don't

need to read everything. We just need to read the most important papers for our research question and our literature review. And then you'll notice after we

review. And then you'll notice after we filtered stuff, we're reading a certain percentage. But then all of these are

percentage. But then all of these are going into mapping concepts and gaps. So

down here is the next important stage and it is probably one of the most enjoyable stages where everything starts to come together. Nice. Okay. Research

gaps and understanding the concepts.

This is where I would use Notebook LM and consensus. So in notebook LM I would

and consensus. So in notebook LM I would be using their mapping feature. I think

it's really really powerful. So if we head over here and we just create a mind map. It'll go away. This is one I

map. It'll go away. This is one I created a while ago. So these are all of the papers I need to know. And you can see it's already structured like a literature review. It's got the concepts

literature review. It's got the concepts we need to know about and this can be used to start thinking about the key themes that we want to talk about in our literature review. So this would be

literature review. So this would be great for introduction challenges of conventional OPV devices. You can see here when I click it gives me a nice little AI summary. And then we've got organic solvents, faith segregation

control. We can click that. Oh, look.

control. We can click that. Oh, look.

There's even deeper. We've got

essentially here some potential headings and subheadings. Yes. And it

gives us a nice little overview of all of the papers we've uploaded. And it

just makes it so so so easy. All right.

The next thing that we need to talk about is consensus. So consensus here, you can see that it gives us this matrix. I really love this for getting

matrix. I really love this for getting to grips with what areas of the research field have been deeply explored and where there could be potentially gaps for you to talk about leading into your

literature reviews. It could be like

literature reviews. It could be like this needs to be explored a little bit more. All right then. So here you can

more. All right then. So here you can see it's got the application and area and how many studies are in each one.

And we can see that therapeutic application of plant systems. There's only one study down here for COVID 19.

Oh no, this one was crisper cast 129. I

don't understand what that means, but uh yeah, you can see here multiplex editing. There's only one paper here. So

editing. There's only one paper here. So

we can get a lay of the land. So we can see that there's loads of studies up here in editing efficiency and there's not many down in this little quadrant.

So we can start saying that the bulk of the research is answering these questions and there's research gaps down here. And if there is a genuine research

here. And if there is a genuine research gap, consensus will say gap. It's easier

than ever to find a research gap. So

that is how I would be using these two tools as part of an AI literature workflow to be like, okay, look, we've got these concepts. We're mapping out our understanding. We've already fully

our understanding. We've already fully read some of the stuff. So now we're really getting to grips with what that literature review could look like. Now

we go into this square. Oh no, the dotted square of doom. No one likes it in this dotted square of doom. It is

horrible. It is full of confusion and understanding and iteration. But that is the suffering we need to go through to get our literature review across the

line. So the next step is structure. Now

line. So the next step is structure. Now

from your understanding of concepts, the mind mapping, the gaps that you found, you've now got some idea of structure.

And I like to put it into chat GPT or whatever large language model you like.

I like to say, okay, I've got a literature review. This is the

literature review. This is the structure. And you can actually just

structure. And you can actually just talk this into chat GPT in some sort of weird random sort of like thought uh explosion out of your mouth and just say, look, these are the things I want

to cover, blah blah blah. And then it will give you a suggested structure if you ask for it. Just speak to it for 10 minutes. have these open and say, "Hey,

minutes. have these open and say, "Hey, I want to talk about this. I want to talk about this. Give me a structure."

And I always work with literature reviews from the broad ideas down to the tiny ones. So, first of all, you want

tiny ones. So, first of all, you want structure. What does your heading stuff

structure. What does your heading stuff look like? What do your subheadings look

look like? What do your subheadings look like? That is all we want from this

like? That is all we want from this structure phase. If you want to get a

structure phase. If you want to get a little bit of a taste of the types of structures you could use, you could use something like thesis AI, which you can

put all of this literature into thesis AI and get a literature review. So this

is what I did here. This is thesis AI and uh you can see I got a really detailed indepth literature review. So I

could look at this and be like, okay, what structure is Thesis AI using? what

type of stuff is it talking about in what order? I can then sort of like

what order? I can then sort of like start to build that academic um understanding of a body of work and say yes this is how people report on that

field and we can use this not to copy and paste across but rather to get an idea of the sorts of structures that exist for talking about this research

field. Look, this is a 36page completely

field. Look, this is a 36page completely referenced in-depth literature review.

Clearly, we can't submit that as our own, but it does give us a fantastic place to start if we're not sure where exactly we want to start our literature review. Okay, then we've got structure

review. Okay, then we've got structure or we can also use cyspace because sci-pace has got an amazing way of uh having a new agent. So, if we go here, you can see that I could put in create a

literature review. Here it is. do a

literature review. Here it is. do a

scientist scientific literature review on blat include papers that are blat this agent will go away. It will create a literature review for you and once again it's not to submit this literature

review because the literature review from a university perspective is for us to show that we understand a particular research area. We can't just copy and

research area. We can't just copy and paste this and be like ah done. You can

do it in like industry where they're concerned about cost-saving, but this is about your understanding and you're cheating yourself if you just copy and paste this across. But what you're really doing is generating all of these

literature reviews and then using your magic excellent academic brain to work out what you do like and what you don't like and incorporating it into your literature review. That is how we use

literature review. That is how we use these AI tools at the moment. But if

you're in industry, you just bloody create one and you go, there we are.

Read it. Done. They're worried about cost-saving in academia. We're worried

about you learning and building up the real kind of uh base academic skills that will serve you for the rest of your career. All right, then. So, let's get

career. All right, then. So, let's get back into the presentation because the next step is content. Content. Donk

donk. There we go. Content. Now, you can cheat yourself out of building skills very easily in the content space now. So

if you're doing a literature review for a university assignment, you cannot use something like Manis or Genpite, this is the Manis logo just to say, "Oh, here's a load of stuff. Create a literature

review because it will do phenomenally well." But you need to make sure that

well." But you need to make sure that you are in control. And so you can use something like chat GPT or you can use Oh, look at this. Like I should have gone full screen.

There we are. We're back here. Full

screen. All right then. So, now we need to make sure that we use uh tools that we're in control. Now, Jenny is in the gray zone. This is Jenny AI here because

gray zone. This is Jenny AI here because as you type, it will vomit out information, but you're still a little bit in control because it's like, "Yeah, I do want that. Yeah, I do want that."

But it is taking away a little bit of the thinking process for you coming up with the narrative for your literature review. It will site as you go along as

review. It will site as you go along as well. So, you can use something like

well. So, you can use something like Jenny AI. Um, but we do have to make

Jenny AI. Um, but we do have to make sure that it is allowed within your university's uh rules and regulations.

All right. And then we've got this.

We've got source uh site. No, sourcely.

That's what it is. Jesus Christ, there's so many of these tools. Their logos all look the same. So, this is a sourcely.

And so, as we're typing, this is actually something that quite often happens is as you're putting together content, you've got the structure and then questions pop into your mind.

you're like, "Oh, I wonder if there's something about that." You can use source lead to go find sources that sort of support some of the things that you're saying that maybe haven't been sort of bubbled up in your first

literature search. So, this is a great

literature search. So, this is a great place in this content sort of box to be like, "Okay, I need to do this a little bit more deeply. I need to understand this area a little bit more deeply." And

you can use something like source or even just go back to this box here, find literature, and then sort of inject it back in down here. But once we've created the structure, we've got content, we need to review it. And there

are two ways that I recommend you can review today. And the first one is AI

review today. And the first one is AI review. So AI review, I recommend you

review. So AI review, I recommend you could use like any large language model here. I've got chat GPT. You can just

here. I've got chat GPT. You can just say, hey, this is the matrix that I need to sort of satisfy or the uh learning outcomes for my literature review. Is it

hitting all of the right notes? And then

it will uh, you know, let you know where you're going wrong essentially. But you

also have got thesisify and paper wizard. So let's head over to something

wizard. So let's head over to something like thesis where you can just sort of like create uh I need to log in.

>> Meanwhile, >> all right, with theify I'm all logged in now. You can actually sort of like put

now. You can actually sort of like put in something this is a paper and it will give you an idea of the summary. to what

works well, what can be improved, over assessment, and you can use this information then to analyze and critique your own work as uh re an AI reviewer has told you that something's wrong with

it. And you don't have to agree with

it. And you don't have to agree with this AI reviewer. You can be like, "No, actually, I quite like the way I've talked about that and I have done what you've said." But this gives you a

you've said." But this gives you a really nice way of being like, "Yes, okay, maybe here suggested topics, purpose. Have has my purpose been

purpose. Have has my purpose been properly understood? Is my thesis

properly understood? Is my thesis statement is it understood? Am I using the appropriate evidence? All of this is bubbled to the top and it makes it very easy to understand where your stuff and

your literature review is falling short.

I absolutely love the fact that that can be done. And another one that can also

be done. And another one that can also do it is paper wizard. But ultimately,

this AI review allows you to take a step back and be like, am I actually satisfying the appropriate kind of part of this literature review? Have I put

forward the right evidence? Have I shown things? Have I cited things the right

things? Have I cited things the right way? This is where these AI reviewers

way? This is where these AI reviewers can really help you polish that work.

Now, after we've done this, what I recommend you do is head to a manual review, which means, oh, you're going to hate hearing this AI tool lovers, where

you read every single sentence, every single word that you are going to submit. And this is where it gets nasty

submit. And this is where it gets nasty and challenging. And also manual review

and challenging. And also manual review can also be done in a peerreview way.

Give it to someone in the field. Give it

to a postoc. Give it to a more senior PhD student and ask if they would wouldn't mind just looking over it and giving you the red pen of doom all over

it. That's what you want. You want

it. That's what you want. You want

feedback. You want to know where the weaknesses are. You want to know where

weaknesses are. You want to know where you've gone wrong. And that is one of the most painful parts of writing a literature review or any sort of part of academic writing is that you have to get

feedback and it is soul destroying sometimes but that is part of the process. You get used to it. You get

process. You get used to it. You get

used to it but initially it's not what you want to hear but then you go back after this manual review and maybe it's something to do with structure. Maybe

it's something to do with content. And I

recommend you just bubble around in here. Blah blah blah. Dart around. Spend

here. Blah blah blah. Dart around. Spend

as much time as you can in this part.

Bouncing around, refining your ideas, making sure the structure makes sense.

You're not putting the cart before the horse, which is what my supervisor used to put all the time on my uh pieces of work. But once we're happy, the last

work. But once we're happy, the last step always in this should be a manual review. And there's two types of manual

review. And there's two types of manual review. The first one is the content. Is

review. The first one is the content. Is

it standing up to the academic sort of uh rigor and presenting the literature of the field in the right way? And the

second way is does it read well and can people understand it? Is it free of typos? For that one, I actually like to

typos? For that one, I actually like to read each paragraph backwards. So start

at the end, read the last paragraph, the second last paragraph because sometimes it's hard to see those silly little mistakes when you're so focused on the narrative. So, doing it backwards means

narrative. So, doing it backwards means you go, "Oh, that sentence doesn't make sense." Or, "Hey, there's a typo there."

sense." Or, "Hey, there's a typo there."

It just allows you to um change your perspective on the literature review so you can see all of the errors. Quite

often artists when they're doing portraits will turn the portrait the other way around to be like, "Oh, I can now see it in a different way and I can see all these errors." That's what you're doing with your work by reading

each paragraph backwards from the back to the front, looking for those silly little mistakes that will just annoy someone and give them that feeling that this literature review was rushed. All

right, then last thing. Once you're

happy with a manual review, something that you're happy with right now, but when you read it in a couple years, you're going to be like, "Ha, that was rubbish." And that's a good thing by the way because then your

skills would have grown and you would have grown as an academic. But be proud of what you created. This is the ultimate workflow for creating an awesome literature review and making

sure you don't rob yourself of those crucial skills that you need in academia and research. If you're worried about

and research. If you're worried about using AI writing as part of your literature review, go check out this video where I talk about using uh AI detectors and how you can get around them or just make sure you're on the

right side of the law, i.e. the

university law. Anyway, go check it out.