Speaker 1: ChatGPT has become so powerful that it can now do loads and loads of things to help you write your peer-reviewed paper drafts. So, first of all, you need to start with figures. It is my process and method to start with figures whenever I write a peer-reviewed paper. It is what needs to be done because it gives you an idea of the sorts of information you're going to include. So that's where this process starts. It starts with this table or this graph, this scanning electron microscope and the AFM data. It starts with this scanning electron microscope and the wavelength transmission study. Starts with this, the Raman study. Starts with this, the bend cycle. Starts with this, the devices that I made and the best device. All of these are important to have before you even start writing a paper because it gives you an idea of the story that you could potentially create. But the first ChatGPT thing that you need to know is that it can actually help you organize what order your figures go in. So here, for example, I've put in all of the different figures that I've just shown you and I've said, hi ChatGPT, I've collected this data during my research and now I'm feeling the pressure to publish. My supervisor is on a sabbatical, which we all know is just a holiday from the assholes in this department. So can you help me by suggesting an order for these figures to craft a, this is an important bit, compelling research story for peer review? And one thing I've found is that if you give ChatGPT a little bit more background information on what you're up to and the state of academia, it actually does a little bit better. So here we are, it says to craft a compelling research story for peer review, you'll want to organize your figures in a way, blah, blah, blah. And it says, how about you start with introduction of materials and methods? Brilliant, that's the schematic. How about that characterization of the materials? Brilliant. How about all of the Raman? How about all of the SEM images? That's what it's telling you. And then how about experimental results, so the JV curve? So should I put the JV curves a little bit earlier on than I was thinking in this paper? This is something that allows you to use your own critical thinking skills and go, you know what? Actually, no, it's better a little bit later, but at least this is telling you and giving you options for how you could lay out that compelling research story. In fact, putting these results a little bit further up may actually be the thing that keeps people engaged and interested in my papers. Discussion of results, conclusion. So it says you might include figures that summarize the overall performance or stability. So like the stress, for example, that's probably a good idea. And then concluded figures support your claims about the significance. This might be the figures that best show the potential improvements or advantages of your novel material. So that could be the best device I've ever produced and I can talk about future applications. I really like it. So use ChatGPT to structure and provide information on the best way to tell your research story. Great. There's more, even better ways that just blew my mind. Check these ones out. Now that you've got an order of figures for your peer-reviewed paper, now it's about getting into the nitty-gritty of what actually each individual figure is all about. So check out this prompt. I put in one of the figures in high resolution. It has to be high resolution so it can actually read all of the details in the figures. But here, I've got this prompt. My supervisor is coming up for a promotion and has become intolerable with their desire for papers to increase their H index. I don't think they're sleeping very much because they're falling asleep in the department seminars and dribbling on themselves. I want them off my case. So can you help me by coming up with talking points for this figure that shows an interwoven network of silver nanowires and single-walled carbon nanotubes for a transparent electrode? Any little bits of information you can give it will help it sort of like reproduce or create really great responses for you. So here we are. Certainly, here's a breakdown of the potential talking points. It goes through each individual panel and it says, how about you just describe the general morphology? How about you just sort of like describe the scale of the inset and what it reveals about the interfacial section between the silver nanowires and single-walled carbon nanotubes? They're talking about this one. That's pretty cool. So it's given you actual talking points. And of course, the talking points is what you're going to expand on a little bit later. And I've got a trick for actually doing that. So stay around because it can even get into the really detailed stuff like the electrical characterization. It knows that one side is the height resolution and one is the electrical characterization. It knows this is height based on the units here, I imagine. And then it knows here is electrical because of the current data here. That's just incredible to me. So here we are. You can now structure and create a story and then under each one of those figures get a list of talking points. And then it is literally about talking. Check this out. This is pretty cool. Now I've got this list. Are we recording? Yes, everything is recording. Okay, I panicked that we weren't recording. Now check this out. Once you've got the individual talking points, you can use a stream of consciousness just to get all of that stuff off your mind. You don't even need to put it into those perfect like academic prose. You can just talk about the figures that you're going to present. And you can use chat GPT to create order from that chaos in your mind. Brilliant. So here we've got the figure that I just showed you. Then I use the talking points just as like a little script and I just let the words flow from my mouth. Ohhh, no order at all. So this is what I said. Oh, by the way, you can go up here and go to voice typing and then you can just click here to speak and it will just record your voice and put it into text. Brilliant. And you can see this is full of errors, but it doesn't matter because ChatGPT is clever enough to work out what you actually mean. It was amazing to me. So here we go. Here's a figure which describes blah, blah, blah, and I actually just talked aloud. In panel B, there is the same silver nanowire network, but the addition of single-walled carbon nanotubes at the same magnification. Down here, you can see in panel D, we could see the silver analyze, hmm, and carbonate tubes, so I didn't say that, but it's good enough, okay? Are electrically connected, which would explain the lower sheet resistance, and interestingly, the silver wire and single-walled carbon nanotube networks, electrically conducted. Do not worry about what it's actually sort of like spilling out, unless it is really, really just quite the word orgy. You liked that last time. Okay, so here we go. This is it. This is what I need to take. And so if I go over to ChatGPT, and I put in the figure again, I think that's very important that it knows the figure that you're talking about, and then I want to create something that's first draft for peer review, so I put it in, and this is the prompt. I've given up on thinking for myself, because all I feel is the impending doom of grant application rejections. It turns the department into a career funeral for everyone other than the professor that gets all the grants anyway. You know the ones. They're in your department, too. He also loves to boast, which is like dancing on the graves of their CVs. Anyway, can you turn this voice transcript into a paragraph for the results section of an academic paper? And then I literally copied and pasted that stream of consciousness, and it's done a pretty good job. So down here it says, figure X presents a comparative analysis of silver nanowire networks transferred to a substrate. Panel A depicts the dense array of single nanowire, blah, blah, blah, and it goes on. It puts on the information. It talks about the panels, and this is a great first draft. You can even use this as a base and then sort of like have another stream of consciousness, paste that in with this, and it gets better and better. I've done it a couple of times. That's how you sort of go now from a story structure to story bullet points, and now to a paragraph under each figure talking about what that figure is showing you. Brilliant, and it doesn't even stop there. This is even more exciting. Once you've got all of your results section, essentially you can go on and you can create an abstract, which is what people read anyway, about all of the stuff you've just talked about. So you do need to provide it with a structure, however. So I've got an old paper of mine, and I've said here's the abstract. So essentially just create something similar, but with the results that I'm going to paste into you. And so here I've got, go to the top. People only read the abstract of a paper before citing it, so can you please turn this results section into a highly citable abstract for a peer-reviewed journal submission? Here is the structure and example of an abstract you can use to structure this one. Ensure the length is the same. So I put in the template I wanted it to use, and then I said, it's a Friday, so I'm hoping to pretend to work for the rest of the day. Here is the results section. Then I put in everything that ChatGPT spit out, and you can see it's a load of stuff. So I'll go all the way down to the bottom, and essentially here it says, innovative, flexible, blah, blah. I was like, okay, it's going well, it's going well, it's going well. And then I was like, ugh, it's too long. ChatGPT, why are you doing this to me? So I stopped it, and I said, make it only one paragraph. I've got to get to the pub. And then it did exactly that. It said, efficient and flexible, blah, blah, blah. And this is such a really great start for an abstract. I'm not saying it's perfect. Don't jump up my ass telling me it's not perfect, because I know, but it is a great starting point, and then you use your magical academic mind to look at that, and then you do more of a reviewing process than a creation process, and the reviewing process is much more interesting, because you can look critically at what's being created and change it. So yeah, do that. We're not saying use this exactly. There's more things you can do. Check this one out. The last thing, once you've got everything together, your results, your methods, your discussions, your conclusions, your abstracts, everything, you can then just put it into ChatGPT and tell it to act like your favorite reviewer. So over here, I've got the PDF with all of the data in there, and I'm like, okay, tell me if this is good enough for peer review, and I've got this magical prompt. Act as a grumpy peer reviewer for a science journal that has been divorced twice and has a high tolerance for alcohol and feeling lonely. Please provide feedback on this article draft, and it looked over this article, and it says, okay, look, this is the feedback as a grumpy peer reviewer, and I'm not saying that this is something you definitely have to change, but once again, it starts you on that thinking. For example, figures. Even though they're well-structured, some figures could benefit from clearer annotations or explanations in the captions. Maybe I should go back and look through and just say, you know what, does everything come across clearly in this figure caption? If not, it's time and opportunity for me to change it. Up here, it says, while informative, it leans heavily on historical context, a more direct approach, highlighting the current challenges. So maybe I started too early in the introduction. Maybe I should just start with the stuff that's immediately applicable to this paper. So that could really wind up a grumpy reviewer that's divorced twice, definitely. And then over here, it's got, the references are comprehensive, but they seem to be over-reliant on older sources. Oh no. So then I can go look for newer sources to put in just to appease this lonely academic's sadness towards this paper. So there we are. There are so many things that you can do and use ChatGPT for. Now with its visual kind of input, now with the ability to put in PDFs and drafts, it is amazing, obviously. Now this is something people always say, what about the data security? I don't know. I don't know. I don't know what it is. I'm sorry. The data security stuff, you know, if you feel comfortable uploading it to here, then brilliant. If not, don't. Don't do it. But I think I am confident enough to put stuff like this in here because this is such a valuable tool, but that will depend on how you feel about it. But let me know in the comments what you think. So there we have it. That's how you use ChatGPT to write a research article. It's amazing what it can do. Let me know in the comments how you've used it, what you're going to do in the future. Now remember, this is just a first draft. You still got to use your sciencey and research brains to make it better, but I've never ever seen something so great at just producing something that in the past, when I was doing my PhD, would take me literally hours. Now it will take an hour. Crazy stuff. This is the world we live in. Oh, look at it. Isn't it joyous? Okay then, remember to go check out this one where I talk about using ChatGPT for peer review. There's more tips and tricks in there and I think you'll love it. So I'll see you over there. Go check it out.
Generate a brief summary highlighting the main points of the transcript.
GenerateGenerate a concise and relevant title for the transcript based on the main themes and content discussed.
GenerateIdentify and highlight the key words or phrases most relevant to the content of the transcript.
GenerateAnalyze the emotional tone of the transcript to determine whether the sentiment is positive, negative, or neutral.
GenerateCreate interactive quizzes based on the content of the transcript to test comprehension or engage users.
GenerateWe’re Ready to Help
Call or Book a Meeting Now