We’re going to focus on AI tools and specifically how to use AI properly as a researcher, and also some use cases where you should be careful when you’re using AI. So hopefully it’s going to be informative and interesting and intriguing and all those other wonderful adjectives that start with I.  So I think we’re ready to start.

 

We don’t need to wait anymore, so let’s start.  We’ll save the Q& A for the end. So if you have any questions, you can just write in the chat, and then hopefully we’ll get to it by the end of the presentation.  Okay, so my name is Abderisak Adam. I did my PhD in civil and environmental engineering  and worked as a researcher for many years.

 

These days, I work specifically as a research coordinator at Avidnote, trying to sort of bridge research and AI tools. Um, and specifically what we look at is how we can use AI in research projects, uh, from writing, reading, analyzing all of the different aspects that are involved in the research process, uh, and how AI can be utilized in sort of an ethical way, um, and so for the talk today, I want to talk a little bit more broadly about AI, um, I’m sure most of you are familiar with AI tools,  um, and so I won’t go into too much detail.

 

So on the right here, we have the most famous example of an LLM, uh, which is Chachapiti. Um, essentially, these language models, all they do is, yes, they try to guess what the next sentence should be, what the next word should be, uh, what the next paragraph should be. Uh, so they’re very similar in that sense to, uh, your mobile phone.

 

So when you’re using your smartphone and you write to your friend, like, hi, Alex, uh, do you want to meet at my, and then it says, you know, either my house, you know, or maybe it says my office, if you don’t have a work life balance going on, you’re always in your office. So, um, the, the, the, your, your phone will suggest my office because that’s what you always write.

 

Or my place or something like that, but the AI or the phone in this case would never say something like my graveyard, even though, technically speaking, that’s grammatically correct. The syntax is correct,  but nobody says it. And because nobody says it, it’s not going to show up anywhere in the data. And statistically speaking, it’s very rare that you would want to say something like that.

 

So the AI is never going to suggest it. So in a sense, it’s just a statistical model to try to understand what the next sentence should be, but because the LLMs have been trained upon, you know, the entire internet, essentially, uh, they can predict not just words, but entire paragraphs and entire chapters.

 

Uh, and so this is useful for us as researchers in many different aspects.  One of these aspects. Deals with reading. So I like to think of it in terms of what do you do as a researcher? Well, one thing you do is you read papers, then you write papers, then you analyze data, and of course you have to design your study.

 

So when we look at it from these sort of perspectives, we can ask ourselves the question, how can we use AI for this specific purpose? Uh, in order to sort of make that process more efficient. When it comes to reading, you can use the AI to simplify. Uh, text. So let’s say, for example, you’re reading something.

 

It’s complicated. It’s convoluted sort of academic jargon. You don’t understand what the author is saying. Uh, and you want to sort of break that down so you can use the AI to simplify concepts. Uh, and easily, quickly get feedback on whatever the person, uh, meant when, when, when they wrote that text. Uh, and we’re going to look at some concrete examples later on inside of the app to show you how that would work.

 

But first, I want to give you the generics of how you can use AI regardless of whatever tool you, you intend to use, uh, in order to just get that framework of how you can use it and how you can’t use it.  Of course, you can use the AI to translate text as well. A common mistake people do when they’re translating text is they tell the AI, you know, translate this from Spanish to English or English to Spanish or whatever.

 

And then you get a translation and maybe you’re not happy with that translation because it’s too direct, it’s too informal or maybe too formal. If you don’t tell the AI exactly how you want the translation to be, it can’t read your mind, right? So we don’t have the chips implanted in our brains, at least not yet.

 

I hope to never have, but you never know. Some people will probably opt for that. Um, so the AI can’t read your mind, which means essentially you have to tell it how you want your translation to be done. So you tell it, for example, I want a direct translation word for word, or you tell it, I want, uh, an idiomatic translation.

 

in which case the AI is going to try to keep those subtle tendencies that exist in that specific language. So for example, in English,  you would say it rains cats and dogs, right? That’s an expression. Um, that doesn’t work in like any other language. So for example, I speak Swedish. If I were to do that word for word translation, it wouldn’t make any sense.

 

It would be like completely nonsensical. So instead I would have to opt for sort of a idiomatic expression that equivalent to that. So maybe it rains heavily or something like that. So it’s important to always tell the AI and modify basically, uh, what type of response that you’re looking for. Uh, the same thing holds true when you’re creating summaries.

 

So imagine you’re reading a paper, uh, or you have data you have to go through and sift through.  A lot of the times people will just copy all of that and then go to something like ChatGPT or a lot of other tools that exist today. So you have Anthropic, you have, actually you have a lot of, a lot of better tools than ChatGPT today that are available specifically for research purposes.

 

Um, and we’re going to talk a little bit more about that later on. Uh, but regardless of which tool you’re using, instead of just copying and pasting in the text, if you modify your response and tell exactly what type of summary you’re looking for, you often get much, much, uh, better outputs. So. If you just tell the AI, give me the summary, you might get a summary that’s like, you know, two sentences long, you might get like a, you know, Reddit style summary that’s like a too long, you didn’t read like one sentence, you might get like a very detailed summary.

 

So if you don’t tell the AI once again, it can’t read your mind. So if you tell it, I want an analytical summary that’s like, you know, roughly half a page long and cover these aspects, then the AI is going to try to cater to those needs. But if you don’t tell it, it doesn’t know, it might, you might get a good, uh, summary, you might get a really bad one.

 

Uh, you can, of course, also use the AI to get opposite views, which I think is a really, really cool sort of feature. Uh, which essentially what it means is that imagine you’re like, maybe you’re writing your literature review, uh, and you’re trying to, uh, present all of the varying, uh, sort of, uh, opinions that exist in a specific topic or regarding a specific topic.

 

Uh, so what you can do is you can just, whenever you read something, you can take that thing that you’re reading and then feed that back into the AI in order to get. What is the opposite view in the academic space for whatever thing this author is arguing? So  maybe this person is saying this is good.

 

Other people may not agree. And so you want the contrasting view, which is what we’re always looking for as academics, as researchers. Of course, doesn’t mean that it’s true or false. It just simply means I need to be able to present all of the different opinions that exist on this matter, uh, in order to sort of situate that within my, my own paper.

 

Um, and then, of course, when you’re writing your, uh, when you want to get an academic critique, you can, you can let the AI be the peer reviewer before you’re peer reviewed, sort of like having a mock trial, uh, before having a, a human, uh, uh, sort of critiquing, you’ll, you’ll have the robot critique, critiquing you, uh, and the AI is really, really good at finding mistakes that are sort of, say, grammatical mistakes, obviously it’s really good, um, syntax related mistakes, inconsistencies, so for example, let’s say you say in one page you say, Uh, refer to table four, you know, for this specific thing, uh, and then the AI can read the context and you see like, oh, did you mean table three, not table four, uh, because it seems like you were meant to refer to table three.

 

So these types of inconsistencies, like the AI is really, really good at finding those.  And of course, other other things as well, which we can take a look at later on  when it comes to the analysis part. Now, this is when you have data that you’ve already collected. So let’s say you have collected quantitative data like from a lab, or maybe if you do interview based studies, you have like interviews that you’ve transcribed and you want to AI to help you out, be your sort of research assistant that comes up with ideas.

 

So these types of ideas are essentially like correlations and insights within the data set itself. So whatever the AI throws at you, obviously you don’t just swallow that. And, you know,  you don’t just, uh, uh, um, accept that without critiquing it. Right. So you’re always looking at.  How? How can I substantiate this?

 

So if the AI comes up with a suggestion like here’s the interpretations of the specific data that you have, and if you look at that, you’re like, this looks interesting. So you do your own analysis on top of that in order to see if this is something that can be substantiated. And of course, because you’re the grown up in the room, the AI can sometimes come up with really nonsensical stuff as well.

 

And you look at that and you’re like, this is complete garbage. I’m not even going to spend the time to, to look into this deeper. So I like to think of it like having a research assistant, that it’s like a grad student, um, you know, grad students who are like, uh, quite new in the field. Uh, they may come up with some really cool things, but they may also just due to inexperience, come up with stuff that’s not irrelevant.

 

So you have to sort of be the grownup in the room and steer the conversation towards something that’s useful. And of course, if you find something that is useful, Double check, make sure that it’s accurate. Um, and then when you are, uh, uh, adding that into your own paper, uh, for transparency purposes, you always state in the research methodology section, how you did your analysis.

 

Uh, and this is not even in terms of something that’s contentious. It’s just the same thing that we did even before AI was a sort of actor. So, for example, if you use in vivo in the past to do analysis, you would say, I used in vivo to do this, or I used Wolfram Mathematica to do this. So it’s the same way you would state whatever software you used in order to do your analysis.

 

Now, if you do interview studies, which I think some of you do, especially for those of you, maybe you do medical studies or maybe from the social side, social sciences. You can do even more because what you can do with the latest tools now is you can upload an interview and you can get a transcript in like in a matter of minutes, you get like the entire transcript, and then you can take that transcript, you can go to the AI, and you can get an analysis of the whole interview with all of the coding, using whatever coding schema that you like, um, uh, and whatever framework you want to base it on.

 

So essentially work that would usually take like hours to do can now be done in literally minutes. Um, and then it’s, of course, up to you to make sure that it’s accurate and to rerun it if something’s not working. Um, and so it’s always as an aid. The AI doesn’t replace you. It’s like an assistant.  And then we come to the methodology part.

 

So this is like when you’re designing your study. Uh, you can use the AI to get ideas on your actual study, the research questions. Uh, if you do interview studies, like in the past example, uh, you can get suggestions on your, you know, interview guide, like a semi structured interview guide. Um, you can do the same thing for people who do questionnaires, uh, or focus groups.

 

You can get the questions that you could ask, and the AI is really, really nifty at coming up with really sort of creative ideas on, on questions that you could potentially pose. And then of course you need to navigate it, discard what is useless, uh, and, and retain what is useful, uh, which is what we do as researchers, uh, right?

 

That’s literally our, our sort of main, uh, task.  And of course you can do more here as well. So you can use the AI as a way to bounce ideas off of. Um, in terms of what type of research method should I use? What type of theoretical framework is useful to, uh, to use for this specific study and so forth.  Uh, and then we come to the final part, which is sort of the big part and which is the part that is also, uh, in, in ways the most exciting, but also in ways where we need to be more, most careful.

 

Uh, and the reason is the AI will and can write original text. Uh, so when the AI is writing, it’s not writing like, for example, when you write something in Google and you search for something, Google will give you a text that somebody else wrote. Basically, all Google does is like a file cabinet. Uh, so it’s copying and showing you whatever somebody else wrote.

 

Whereas with the AI, with these LLMs, what’s happening is that they’re writing original text. Um, so if I were to ask anybody, everybody in this room, this sort of virtual room, I want you to write a history about the United States.  You would all write your own text. Assuming you didn’t collude in some way, you would all write original text that is completely different from one another.

 

Uh, however, you would touch on the same points, you know, 1776, you know, the, the, uh, the, the, the, the, the war, the internal war that occurred later on, you would, you would, you know, II, you would talk about like major events Um, that affected the country. Um, but you wouldn’t write the same text. So it’s the same way with AI.

 

It’s sort of having like a person who has multiple, you know, personality disorder. Every time you put an input to the AI, it will give you a new output. And the reason for that is because these models are non deterministic. Non deterministic, which essentially means They don’t follow a linear path. They will always, because it’s a statistical model, they will always give you a new output, even with the same input,  which is useful, uh, because if you’re not happy with the output, with the output that you get, you can always just modify the input and you’ll get a new output.

 

Um, however, just because it’s original text that’s being given for the most part, that doesn’t mean that you can just take that text and claim that as your own, obviously you still have to keep academic integrity, which means If the AI comes up with a suggestion of something you can write, you treat that as a draft of something that sort of like ideas that you could pursue.

 

You may just use like maybe one sentence or two sentences or maybe one paragraph, and then that paragraph that you use, you incorporate that within your own text, you rewrite it in your own voice, you discard what is useless, you retain what is useful, but once again, you incorporate that and weave it in within your own body of text.

 

You source real authors. Um, one of the main problems that AI tools have, uh, is the, they don’t always cite sources properly. So you have to be extra careful when it comes to citations. Uh, even though the text will often be correct, the citations will not be correct. So, if we take the example previously, uh, the previous example,  if you tell the AI to write a history of the United States, it’s gonna write a really good, uh, history.

 

Uh, but if you tell it, give me the citations, Some of them will be good, some of the references will be bad. So you have to be careful when it comes to that specific aspect of it. Um, but you can, of course, also use it for other purposes. So you don’t need to write from scratch. You can use the AI also to simply rewrite text.

 

So let’s say, for example, you have text that you have written already, and all you want to do is sort of rewrite it. Maybe you wrote something in an earlier paper, like a conference paper or a working paper, and you want to reuse the introduction or something  for your new paper. So you take the old paper, you copy that, and then you paste it into the AI, and it will rewrite it for you, keeping the same meaning, uh, and then, of course, you make sure that that fits within your new paper, uh, and anything that’s bad, you fix and, and, uh, you know, you improve upon the text.

 

So it’s an easier way to sort of work with material.  You can use it also when you want to cite somebody else. So let’s say, for example, you’re citing somebody. Uh, Stevenson from 2020, uh, Stevenson argues that yada. So you go to whatever Stevenson said, you copy their text, you go to the AI, you tell it to rewrite it.

 

Then you, you, you then modify that within your own text. And you say, according to Stevenson, they are Stevenson et al, they argue that, and then you put in whatever the rewritten text of their argument. And you make sure, of course, that that reflects what they actually said, right? Uh, which is what you normally do when you cite somebody.

 

You rewrite it in your own words, and then you put the citation back to the real author, right? So it’s the same way. So the AI is essentially saving you time in, in finding those ways of phrasing and, and rephrasing text. Uh, but at the, at, at the end of the day, you, you have the re responsibility to make sure that it’s accurate and, and, uh, correct.

 

Then of course, you can use the AI to proofread the text. So you have the three different levels, like the first level is like completely original taste right from scratch. Um, and you may only use like a few sentences. You may use a paragraph. You may not use anything. You may completely, you know, use a new generation because the first one was bad.

 

Um, in the rewritten part, it’s even less original. You want to keep the text. You just want to rewrite it. Maybe you’re not happy with the grammar or the syntax or whatever. You just want to rewrite the text, but keep the meaning. Now in the third part, it’s even less original because here you don’t want to change anything.

 

You just want to fix the mistakes. So it’s like having a polished work, you’re complete, like the, your paper is complete or your thesis is complete or whatever you’re writing is complete. All you want to do is proofread, find mistakes, grammatical mistakes, maybe spelling mistakes, um, and keep the, keep the general meaning.

 

Okay, so with this being said, uh, I now want to jump in to show you, uh, one specific tool that is developed for researchers, by researchers. Uh, which is called Avidote. It’s a web based tool, uh, that essentially uses AI for all of these different purposes. So you can use it for reading, which is what we talked about earlier, about interpreting text, about peer reviewing, simplifying concepts.

 

Uh, you can use it in the writing process as well as a writing aid, uh, for proofreading text, but also for writing original text, which, of course, you, you then have to incorporate your own text and, and, uh, modify the text and make sure that it fits within your own, uh, style of text.  But the AI can come up with ideas on things that you could write, and then of course you can also use it for analysis purposes, as well as when you’re designing your study.

 

So before I jump into the app, I’ll show you just quickly, uh, let’s brief videos of how it looks like. So you can use it in the reading part to chat with the papers that you are reading. So on the right side, I have a PDF that I’m reading a paper. On the left side, I’m chatting using AI. And because it’s connected to the PDF, it will give you concrete, uh, answers based on what’s actually in the paper.

 

And then you can save those answers as notes. You can create your own notes if you want to. On the left side, Um, and those notes can be expanded. So whenever you write notes, you always have them on the left side. Uh, you can, uh, of course change the font and stuff like that, or rather change like the, uh, make it like, uh, emphasize the text or change the colors and so forth on the right side, you can annotate.

 

So you can colorize, you can create post it notes that are sort of, um, appended to that specific paper. And then if you open up the annotations bar, you will see all of your sticky notes, you’ll see all of your annotations, and you can skip between the different sections.  Now, if you want to use the AI to write, you can go to the writing module, the AI writer, you can tell whatever you want to write about.

 

In this case, it’s renewable energy sources. Um, and it’s a literature review that I chose. So it’s going to write the literature review based on that specific topic. I can also state like what type of style style guide that I want, like APA guide, for example. And then you can work with AI in real time to modify the text, to translate text, to expand text or to contract text.

 

So it’s a very useful way of working with your own text as you are writing it. So maybe you’re not happy with like a specific sentence, so you can highlight that sentence and then you can try to regenerate that sentence and regenerate it until you find a version that fits what you’re trying to say, right?

 

Which is essentially what you do when you write. You’re trying to figure out ideas on how to connect different sentences that you are writing. Um, and you can then download that as a Word file.  For analyzing data, you can copy any data in table form from Excel, from other places, and then you can ask questions to the data set.

 

You can analyze the data so you can get, uh, for example, uh, what is being said. Is there any, are there any correlations within this data set? Uh, insights that can be extracted. So we’ll look into this as well, uh, in a few minutes.  Um, and then of course you can transcribe, which is what I mentioned earlier.

 

So if you have, if you do interview studies, you can upload your interviews. It will transcribe the whole thing for you. You can then copy that transcript and you can go to the data and then analyze that interview. And when you analyze the interview, you get all of the different themes, you know, the quotes that are connected in sort of within those themes.

 

And it can save you hours of work in basically analyzing interviews.  And of course, before we jump into the app, I’ll just show quickly these two last slides.  For those of you who prefer the ChatGPT module where you’re chatting with the AI, you can of course have that as well. So if you prefer that interface where you’re chatting without the PDF, you can do that as well.

 

The key difference is all of the data here stays private. So unlike in chat GPT, where it’s not private by default here, everything’s private and it’s never, never used to train the AI. Um, because we have to be compliant to both data policies within the EU where we’re based as well as, uh, data regulations in the U S and, uh, and Australia.

 

Um, and then finally you can manage everything in Abenote from one place. So it’s sort of like comprehensive suite of tools for using AI, where you can store all of your papers, notes, and references. You can find new papers, you can upload your papers, and you can organize everything using tags.  Okay, so I think we can jump in now, so I won’t, um, take too much time.

 

So we can jump into the app, and then we’ll do the Q& A at the end. And I’d love to hear your thoughts about what you think about all of this. Uh, so now I’m inside of Avinote, so if you want to follow along, you can, you can, uh, register. Um, in, uh, you go to avinote. com. And you can create a free account. Uh, and once you register, you’ll come to this view here.

 

Um, here you’ll find under read, you’ll find like all of the functionalities that you would reading under writing, you’ll find everything that deals with writing and so forth. So if we click on my library, I’ll see all of the papers that I have.  So here as an example, I have a research project that I called CSU.

 

Uh, I’m writing about diabetes research, uh, here. So I can click on any paper. I can filter by the tax that I have created. Um, if I don’t want to have the folder view, I can click on files and then I’ll see every file that I have inside of Avinote. I’ll see all of my tags here that I have created. And when I highlight, when I click on those tags, I’ll see basically, um, every, every paper that has the specific tag with the corresponding notes that I have created.

 

Basically, um, so let me deselect these. Um, if I want to search for a specific keyword, so imagine I have hundreds of papers that I have read over the years. Um, so let me deselect these. Um, if I want to search for a specific keyword, so imagine I have hundreds of papers that I have read over the years.  And I want to find a specific keyword, so I’ll just search that keyword and it will show me every incident where this keyword occurs inside of the text.

 

So, for example, it occurs here in this paper, it occurs again. And if I click on any of these papers, I can just click on a paper. It will then take me to that specific paper.  And here you’ll see my notes that I have created here on the right side. I can highlight text here. I can change the colors and so forth.

 

I can create smaller sneaky notes. Thanks.  do not use, not relevant. Uh, and of course I can change the formatting of this as well if I like to do that. Uh, and all of these annotations will be shown. So if you want to skip between them, you can do that and you’ll, you’ll see all of your different annotations essentially.

 

You can go into full screen mode if you prefer to have full screen mode and simply read without writing. Uh, if you want to use the AI, you can just write. So what is this paper about as an example?  Um, and then the AI will start to answer based on whatever, uh, question that you ask. So I said, what is this paper about?

 

The AI said, this paper discusses total quality management concepts and their implications in the construction industry. It references industry surveys conducted in the U S uh, and publishes research to explore the topic. So you can ask any specific question, uh, who, uh, were the authors of this paper, uh, and then the AI will try to answer based on whatever information is there.

 

So you can see it gives me accurate, uh, information. Um, now of course, uh, it’s an AI, so it’s not a hundred percent all the time accurate, but it’s accurate the vast majority of the times. And that is because we’ve connected the paper with the AI, so it gives you much, much more accurate responses. Okay.  Now, if you want to ask questions that you always ask, like, for example, results, instead of writing, what are the results of this study?

 

Every time you want to ask, you know, and you’re reading a new paper in Avidnote, all you do is just click on results and then it will give you the results as bullet points. And so you can just look through the results and see if it’s interesting to you. Same thing when it comes to methodology or a summary.

 

So if you want a quick summary of the paper, it will give you a short summary. And then because some people want the detailed summary, um, it will give you that as well. So it will first give you a short summary, and then it will give you a detailed summary, uh, with an analysis of,  sorry, of what was said, um,  or what the author basically said in that specific paper.

 

Um, and then of course, if you want to take notes by freehand like this, you can do that as well. You simply click on this tool here, and then you can. Take whatever notes you want by hand. It’s useful, I think, if you have a device where that makes sense, right? Yeah, I’m sitting on the computer, so that wouldn’t make so much sense.

 

It would be better to highlight just by hand in this specific example. You can, of course, download this. You can search for any keywords inside of the  Avidnote View here. And if you like any of the responses, you can save them as a note. Just click Save as Note, and then they will be saved here. So, summarize this document was saved.

 

And, of course, you can alter this. You can, oh, sorry.  I clicked, uh,  so you can alter this and you can change whatever text you want and so forth. Uh, you can create a whole new note if you want, just click add new note and say my brilliant, uh, note,  uh, very brilliant  and, uh, not sure how brilliant that was, but you get the point.

 

So you can, you see it will be created here and then you can see all of the notes that you have created, uh, and you can see all of this chat outputs here. Of course, you can clear this if you don’t want to keep it. Now, in, in some use cases, you may not want to chat with the whole paper. Sometimes you, all you want is maybe to explain one paragraph of text.

 

So you click explain this text.  And what you can do is you can just copy any text that you have. So, uh, let me actually refresh this page. Um,  okay. So what you can do is you can copy any text that you have. I’m going to copy this, uh, paragraph here. And I’m going to choose explain this text and paste it in and click send.

 

And then the AI gives you an explanation. So let’s imagine you didn’t understand what the author meant here by, for example, this topic TQM, what does that mean? Um, so the AI explains the whole text that you selected, um, in easy English. So not convoluted, uh, achieving good quality in the construction industry has been a challenge for a long time, each year, a lot of time, money and resources, and then it keeps on going and then it gives you a glossary of all of the terms that were, that were used.

 

So, for example, TQM, what does that mean? Um, it’s a set of practices focused on, uh, continuous improvement in all aspects of an organization’s operations. Uh, and then it gives you a warning, you should always double check specific citations and so forth, because it’s important as researchers that everything we do is accurate.

 

Okay, so this is how you would use it for reading. Uh, if you want to upload, uh, a document, you simply click Upload Paper, and then it will, uh, you can choose which folder you want to upload to, or just upload it, uh, to the root folder.  Select any PDF, PowerPoint. So you can use PowerPoint also. Uh, here’s an example.

 

So this is a computer science course that I’m teaching. Um, and here are some lecture slides. So if you have slides that you want to, um, analyze or whatever reason, take notes on, you can do that also. Uh, the same thing is true. If you have, for example, a PhD student that you are supervising, you can create like a folder for all of their stuff, and then you can see all of the notes and their progression of their work and stuff.

 

So it’s a much easier way to organize all. research related stuff.  Okay, I won’t dwell in too much detail on that.  Now, if you want to use, um, the, uh, the AI to analyze data, what you can do is you can click on the AI snippets, uh, and here you can use the AI for a variety of different use cases. So you can use it, for example, everything from planning your lessons, uh, for people who have teaching.

 

So let’s say I’m, I’m writing about linear algebra. I’m teaching that to, sorry, I’m teaching that to undergrad, undergraduate students. Um, and so you can plan your lessons. Now, of course, I would be more detailed. If you have a specific topic, it would be better to write the exact topic, and then the A. I will come up with solutions.

 

That’s all right. Problems you can give to your students. Potential assignments, um, you know, uh, presentations that you could do with the answers to those  problems. It’s a very, very useful way of planning your lessons, and all of this then can be saved. So my lesson on linear algebra as an example.  Uh, so the way it works is you, you always, uh, put your input here on the left side, and then on the right side, you get the output.

 

And this you can change. So if you don’t like it here, you can, you can, you can, you can write your own text here and you can modify this, you know, so, um, yeah, whatever, whatever you want to write, you can write basically, um, if you want to use it to analyze data, you would select data. So let’s say, for example, I’m reading this data here, this table,  and, uh, it’s like all of the big companies that exist in the world,  uh, and, uh, where they are based, how much revenue they have, uh, which industry they’re, they’re a part of.

 

Okay. So I’m going to copy this data and I’m going to go to analyze data  and I’m going to paste it in and then I can ask questions to the data set if I want to, I can keep it blank if I just want the analysis, but let’s say I want to ask specific questions. So let’s say, for example, which, uh, or, uh, which, uh, country has the most companies, uh, featured,  uh, in,  in this list.

 

And the other one, maybe we’ll ask, uh, which company.  In the financial sector, uh, had the highest revenue.  Uh, and then maybe what was the revenue of Ford just to see if it can get that as well. So I click generate texts. And so now the AI is going to look at whatever data you inputted here on the left side.

 

Uh, and number one, try to find insights within that data without resorting to anything outside of the data itself. That’s very important. Uh, so in the first step, it looks at the data itself. Okay. And it tries to come up with, um, analysis that is based just on what you put in. So, for example, it says that there’s a dominance, uh, the dominance of the United States.

 

The U. S. is home to the largest number of companies in the list with 14 entries showcasing the country’s strong economic presence, particularly, uh, in the retail and healthcare sectors. It talks about discrepancies in the profitability. Despite high revenue, some companies report significantly lower profits, such as Walmart and Amazon.

 

Uh, this discrepancy suggests high operational costs. Or competitive pricing strategies aimed at market capture. So that’s really, really interesting. It comes up with stuff just based on the data itself that it could be useful when you want to analyze it further.  And then in the second step, it then tries to assess the data.

 

Like it looks at things that are outside of the data itself in order to make sense of this. Like for example, it talks about how the decline in revenue for oil and gas companies might be reflective of a broader trend towards sustainable energy solutions. And fluctuating oil prices, which have affected profitability in the sector.

 

So that’s not inside of this data, like nowhere in that data table, like this table, did it say anything about sustainable energy, but the AI is able to connect that with the broader knowledge that it has, uh, and see that this could be a potential connection that you might want to look into.  Um, and then you can, the AI will give you further ways to analyze the data.

 

So for example, you could do a comparative analysis. Uh, you could compare like revenue and profit margin across various, uh, industries. You could do geographical trends, you could do sector specific deep dives. So the AI comes up with  ideas that you could sort of take further. Like forget about what the AI tells you here, like what, what other types of analysis could be relevant for this specific data set.

 

So it’s like a very useful way of probing your data, coming up with ideas on how to analyze your data, new ways to sort of shed light on it. And then at the end, it gives you answers based on your specific questions. We asked three specific questions. We asked which country had the most companies here.

 

And of course it’s the U S of a, it’s the United States of America. Um, and then it’s asked the second question was like, which company in the financial sector had the highest revenue?  And that’s also true. It’s, uh, it’s, I’m sorry. That’s also an American company is of course, Berkshire Hathaway. Uh, and if you look here in the companies, we go down, we’ll see that Berkshire.

 

Uh, the way it should be pretty high up in the list somewhere, uh, it’s number 10. So in the financial sector, it’s the highest one.  And then the final one was like, uh, which revenue did Ford have? And then you can see the revenue numbers here. And if you scroll down and you look at Ford, oh, sorry, uh, it’s one 76.

 

Uh, and that’s also what you see here. So it gives you accurate information based on, uh, what, what type of data you inputted. But of course, it’s an AI. It’s a probabilistic tool. It’s not accurate 100 percent of the time, but it’s accurate the vast majority of the time, depending on what type of data you input.

 

Um, okay. So if you want to save this, you can save it as we saw before, analyze data. Um, and I’m going to save it under a specific workbook  called data. You can create your workbook, so you can name them whatever you want. You just go to save documents and then you create a workbook there. Okay. Um, and then it will be saved inside of Avid Nodes.

 

So if you open your saved documents, you’ll see all of the stuff that you’ve inputted here. So if I click data here, I’ll see, you know, save data here, data revenue, analyze data. So I can see everything that I have created in the past and I can download those files.  Of course, if I want to work inside of  my computer, or if I want to work inside of Word rather, I’m guessing you’re on your computer in both cases.

 

But if you want to use it, use Word, you can download it as a Word file or a PDF file, and then you can keep on editing inside of Word. So I’m sure you know how Word works, so I won’t go into detail here.  Okay, so this is how you use it to analyze data. Uh, if you have other forms of data, so let’s say, for example, you have, uh, transcription data, you have an interview, so you upload your interview file, so let me go, go and find an interview, um, and then you select the language, English in this case, uh, you select the workbook if you want to, uh, and then you can, if you want to, you can select any of these options, so these options, essentially what they will do is they will label the speakers, So imagine you have, um, you know, person A and person B speaking, it will label the different speakers and timestamps will give you timestamps.

 

Um, some people like to have timestamps. Some people don’t like to have it. I’ve noticed that in the medical industry, it’s very, uh, or sector is very popular to have timestamps. Uh, we have a lot of, uh, users from hospitals and stuff, and they, they love timestamps. Um, and all of the data here stays private.

 

So, uh, it’s never used to, to train the AI. It’s never shown to anybody else unless you show it to somebody else. So it’s all private, and this is because we have to be compliant with all of these different, uh, ethical, uh, guidelines that exist when it comes to, to research data.  Okay. So once the AI has, has transcribed.

 

And by the way, for people who don’t have interview studies, you can still use this for your meetings. So we have people who use this for their research meetings. Um, so let’s say you have your team meeting or you have a meeting with your supervisor or PhD student, then you just want to get a summary of what you said.

 

You can transcribe it and then you can put it into a template that we’ll look at later on. And it will actually give you a summary of everything that was said in the meeting, right? And who said what and who is responsible for what next time and stuff like that.  Uh, it’s very useful for that, for those purposes also.

 

Okay, so now we have the interview. This is completely transcribed, as you can see, um,  and, uh, if you, if we, if we listen to it, we can, we can also hear it. So let me share the sound. So this is an interview with Dr. Dr. Oppenheimer, the famous physicist, Robert, Robert Oppenheimer, you know, uh, from the Manhattan Project, uh, most famous from the movie they made about him recently.

 

And, uh, it’s like from the 1950s or so, the interview, uh, I’ll just play it a little bit. Hopefully it’s not too loud. graduates, because I’m afraid I’d confuse them. I’d have a dozen or twenty men that I’d work with into So this is, uh, like, oh, sorry, let me close this. Uh, so this interview is essentially, it’s, it’s like an interview between him and somebody who wants to ask him about, uh, what work he was doing in his institute.

 

And he talks about like what they were working on. So he says, for example, well, it’s not, it is for many people. There are no telephones ringing and you don’t have to go to committee meetings and you don’t have to meet classes. It sounds pretty nice. Right.  So it’s like even back then people were complaining about teaching like researchers have always been the same.

 

Um, so anyway, he’s talking about all of this stuff and you can see it’s a speaker A and speaker B. And if you click on the timestamp, it will actually  it will jump to that specific corresponding point here. So I would always recommend listen through the whole thing once just to make sure that it’s accurate, right?

 

And then when you’re finished, you can just copy this transcript. If you want to label, like, who is actually speaking, you can, I would say, open up a Word file as an example. Uh, let me open up, like, a Word file. And you can paste it in here. Uh, and if you want to label, you can do that. So let’s say the person who’s asking the questions, Speaker A, let’s say that’s me, Adam.

 

Uh,  and I’m going to replace all.  So in order to open this, you just click Ctrl G inside of Word. Uh, and then Speaker B, that’s, uh, our professor. So that’s Oppenheimer.  Uh, and then we’re going to click Replace All.  And so you can see now the transcript is, uh, now giving me, like, who is speaking with their names.

 

But if you want to keep it private, just leave it Speaker A, Speaker B. Uh, even more so if you want to be, like, super private. You can actually select this one, which will remove all names. It will redact all personal names. So if it’s very, very, very sensitive data, like with names and stuff, just click the PI redaction, it will redact all names.

 

Uh, so that’s for people who have extremely sensitive data. But even if you don’t select it, all of the data stays private, regardless.  Okay, so let’s copy this now, and let’s go to the AI snippets. Let’s go to data, and now we can actually code this interview. So we can code it with like, um, um, uh, a schema that we choose.

 

We can use grounded theory, uh, we can do like a narrative analysis to analyze the interview, which I’m going to do. Um, so we do, we just copy our transcript. We paste it in here on the left side, and then the AI will give us all of the narratives, like what was said in the interview analysis of those sayings, uh, the relevant quotes beneath each section.

 

So for example, it says here that.  Uh, Oppenheimer said, uh, um,  uh, the responsibility of scientists in a technol technological world. And this narrative Oppenheimer reflects on the ethical responsibility of, of scientists in a world increasingly dominated by their discoveries. He acknowledges the ignorance shared by both scientists, and then it keeps on going.

 

And then here you have relevant quotes. So this is an actual quote from the interview. Uh, the problem of a coherent civilization is the problem of living with ignorance and not being frustrated by it.  . Yeah, that’s, that’s, uh, still accurate, I think. That reflects Twitter. Um,  uh, yeah. So essentially what you have here is like all of the different quotes, everything that he talked about with, with specific, like key quotes where, where it’s like sort of encapsulate encapsulates  that specific narrative.

 

So for example, he’s Oppenheimer says here, we are here to take away from men, the cares, the pleasures that are their normal excuse for not following the rugged road of their own life and need and destiny.  People used to be so poetic in the fifties. I don’t know what happened. Um, anyway, so you have the text here.

 

Now, if you want to save it just like before you save it in Avidnote, or you can download the word file. So I won’t talk too much about this. We don’t have so much time left. I want to leave some floors for the, for the Q and a,  uh, but I want to jump into, uh, uh, just one more thing here to show. I won’t be able to show everything here.

 

So if you want to play around with this, I would recommend just create a free account and test it out. Uh, so for example, in the methodology part, you can identify research gaps. You can suggest a hypothesis. You can suggest the framework. Uh, on the right, we’ll talk about this separately in a second, uh, under read, you can critique text, you can proofread text, you can get the opposite view.

 

So imagine you’re reading something, uh, let’s say I’m reading this text here, uh, which talks about, um, uh, nuclear power. Uh, and it’s saying that nuclear power is like a good energy source, uh, it reduces CO2 emissions and so forth. Imagine I’m writing a paper about this and I’m like, hmm,  I wonder what other people say, like, I’m sure not everybody agrees with this.

 

So what do other academics say in the, uh, in the research, uh, or researchers say? So alternative perspectives, the AI will give you whatever the alternative perspectives are in academia. So in this example, these people here are arguing in favor of nuclear energy. So what are the opposite views?

 

Alternative perspective on nuclear energy often highlights concerns regarding safety, environmental impact, economic viability. Critics argue that the potential for  catastrophic accidents. Um, as seen in Chernobyl and Fukushima, and then it keeps on talking, right?  So maybe you discover something new here you didn’t  think about before.

 

So you go back to Google Scholar, you go to PubMed or whatever database you use, and you try to find real papers that are arguing that, and then you interview that within your own writing. So,  use the AI as a idea generation tool and not as a replacement of you. It doesn’t replace you. It’s no more different than using a computer or using, for example, before this, we would use Word to, to spell check our document and give us like the red squiggly line.

 

That’s technically an AI doing that. I mean, I wouldn’t call it AI because it’s like a linear model. It’s not a non deterministic model, but it’s essentially computer code that is giving you all of the grammar mistakes, right? So the AI here can do that and even more. So it’s like just having a tool. You can use a tool in a good way or a bad way.

 

Okay. Um, and hopefully I’ve shown some useful use cases here so far.  And finally, we get to the writing part. So this is the last part. And then we’ll jump into the app. Sorry, we’ll, we’ll finish up and we’ll do the Q and a. Uh, so what you can do here is you can select like what you want the AI to write about.

 

So let’s say I want the AI to write about, uh, I don’t know. Uh, substance abuse maybe.  Um, and, uh, mental health.  And let’s say I want to use the citation style, uh, APA, and I want to generate text.  So now the AI is going to try to generate like a literature review because that’s what I selected and it’s going to try to stick to this topic here that I inputted.

 

The more detailed you are, the better. Um, and you can experiment. If the output is bad, try to change the input. So always just It’s like a different tool than what we used to in the past, right? Because a calculator will always give you, uh, like, uh, the same answer, whereas the AI will give you a different answer all the time.

 

So you have to sort of work with it until you get something that’s useful. Um, so here it starts to write a literature review. You can see the citations and stuff here at the bottom. Um, these are much, much better now. In the earlier versions that we had, they were, like, really poor. Uh, but the version now can give references that are much, much more accurate.

 

However, they still don’t work. mess up. So this, I would still say is the biggest weakness, weakness of AI is that it will, uh, sometimes create, make mistakes. Uh, sometimes it will make up like an entire reference. Uh, sometimes it will give you the reference, but it’s a real reference. Um, so if I were to search for this, for example, let’s see if this is a re reference.

 

Cognitive theory for depression. Okay, seems to be a real reference,  but essentially it will give you like, uh, uh, the, the, the reference, but sometimes it will mess up the name or the author or the year. So you always have to be careful here and make sure that it’s accurate. Whereas the information that you, that it actually gives you is often correct.

 

Like the information will be correct, but it will mess up like the page number or the reference number. So anyway, be careful. Uh, and then if you want, if you find something that is useful, so let’s say, for example, barriers of treatment. So this part is interesting here about the fragmentation of services in another critical barrier.

 

Many individuals with  co occurring disorders navigate multiple treatment systems. This is actually pretty interesting. So let’s say I want to select this and I want to write more about that. So I’ll just click on the AI tool and I’ll click on expand content  and then the AI will write more about this.

 

Um, it takes a few seconds. As you can see, now we have like a much bigger text that continues on on that specific topic.  Uh, you can go the other way around. So you can also summarize. So you can click summarize if you want to shrink content, you can shrink it. Uh, you can, you can improve writing. So let’s say for example, I have spelling mistakes.

 

Um, I’m going to commit a few mistakes there, add an extra the, maybe I’ll add like a spelling here and there. And let’s say I want to select this now, and I want to fix this text, click check grammar, and the AI will fix it. So you can see, it removed the, it fixed, so this is like a capital letter, it removed the extra the, it fixed this here.

 

So, it keeps the same text, but just fixed the grammatical mistakes, the spelling mistakes.  Very, very useful, even for your own text. You don’t even have to generate text. You can just copy and paste in something you’re writing with, uh, writing on, and then you can use the AI to, to sort of, um, uh, either fix mistakes or come up with alternative ways of writing.

 

So if you click improve writing, it’s going to come up with ways to, to change the writing and try to improve the writing. So it’s, it’s very, very useful in sort of bouncing ideas off of the AI back and forth, back and forth. Um, And you can do even more. So let’s say, for example, you want to translate, you can do that also.

 

So if I go here and I try to translate this to some other language, I can select whatever language I want to translate to. And then the AI will translate that specific section to that other language. In this case, I chose Spanish, but we support a bunch of languages. Same thing holds true here as well. So you have a bunch of languages that you can choose between.

 

Um,  and if you want to save it, you save it. So this is about mental health paper. And, uh, you can select the workbook just like before. Maybe I’ll save this under research papers and then it will be saved. It’s, it saves like every few seconds automatically. Um, and you can download this as well if you want to download it, or just copy it if you want to copy and paste it somewhere else.

 

You can do that also. Uh, and then finally, uh, if you want to append a prompt to the whole thing, so let’s say you want to do a custom prompt, your own prompt. Yes, right here under custom action, for example, list all authors mentioned.  And you can see the AI now gave me all of the authors that were mentioned in this paper.

 

So I, I assigned this prompt to the whole text. And of course you can, you can do it in whatever the way you want. You can say, for example, uh, Give me the authors, the results, and the methodology, and give me that as a CSV file. And then it will give you a CSV file, and you can put that into Excel. So it’s an extremely useful way to extract information from  papers or text.

 

If you want to go back, you just go back. So you can use the arrows here to go back and forth. And I won’t go into all of these details here. You can experiment with that in your own time.  If I go back here now, the final thing is, if you do programming, you can use the programming here, select which language you want, and you can get computer code.

 

I won’t show that because I know it’s just a minority of researchers that program. Uh, but if you want to do that, you can do that. If you want to use the same ChatGPT experience, we’re using the ChatGPT API here, the enterprise version, which means that all of the data stays private here. So if you like ChatGPT and you want to continue using it inside of Avidnote, you can actually do that.

 

So, I don’t know, what is the capital of Denmark or whatever?  And you’ll get the same chat experience that you have in, or similar experience as ChatGPT. But the difference here is that everything stays private, and you can keep on the conversation as normal.  And then finally, finally, finally,  if you want to  test this new feature that we have, it’s a beta feature, so it’s still new.

 

It will allow you to find new papers within the app, so you can search for things here.  Like you might as an example, I’m searching for papers.  And then if there’s a PDF available, you can click on the PDF. It still gives you, sometimes it gives you errors. We’re working on fixing that, which is why it’s still a new feature.

 

We haven’t released it, um, but this time it worked, but sometimes you might get an error, which, uh, it’s usually enough to just refresh the page or, or try again and see if the paper was actually uploaded.  Um, and then it will upload the paper, but the normal way of uploading is just to click upload paper, right?

 

So if you do that,  it will, it should work like a hundred percent of the time. So yes, click upload paper, and then it will upload that paper for you from your computer. So. If I go here, I click demo, I upload a demo article.  I open that paper, and it’s going to upload that paper, right?  It takes a few seconds for the AI to process before you can chat, and then once it’s green, you can start using it.

 

Okay, so I think we can finish up. Um,  there was actually a few more functionalities, like you can, for example, export your references or import your references, stuff like that. But I think those are the main features, so you can explore those on your own if you’d like to.  So how should we look at AI? Just very, very briefly.

 

AI is a helpful tool. It’s a helpful tool. And tool means that you use it, right? Which means that you have to discard stuff that is not useful. You keep the stuff that’s useful, but you incorporate your own text. You rewrite it in your own voice. You make it, you make sure that you, you work with the material.

 

So this is not like finding, like, imagine you go to, like, somebody sends you a Word file. Like, here’s a, here’s a draft that that’s finished. No, this is more like Google. When you Google something, you don’t just copy whatever you found, right? Even if that’s, um, useful, uh, you work with that material. Maybe you’ll quote like one paper or one section of that paper.

 

And then you, you, you give the proper source, uh, in the, in the AI example, you don’t cite the AI, obviously the AI is a tool and you don’t cite it. However, you state that you, if you use the AI for generating text, you go to the research method section and you stay there, that’s the typical requirement. Now for most journals is that you stated in the research methodology section.

 

And then you never trust the AI, right? We don’t even trust other researchers. We make sure that what they said was accurate before we cite them, right?  The AI is like an even greater level of sort of critical  approach that we need to take basically when we see what it writes.  So don’t trust. Verify. And then we can use the AI to automate all of this mundane nonsense.

 

Like, why should I transcribe if the AI can do it for me? Like with 95 percent accuracy, it’s better for me that the AI transcribe, and then I’ll listen through it and fix the mistakes. So I save like four hours of work, um, instead of like, you know, doing all of this mundane stuff and I can focus on the actual hard problems.

 

Right.  Um, yeah,  all of the data stays private inside of Avidnotes. Uh, if you want to use it, regardless on which plan you are, you’re using the free plan or the paid plans, there is always a private.  There are no ads on the platform. Um, so the basically the plan is that people upgrade and if they don’t want to use, they stop using it.

 

Um, and if you want to try it out on your, for your group, we give the trials. So if you have a research group for your department or research group, um, just contact me and then we can give you a trial. So you can try out the premium plan also for free and there’s no commitment. So if you don’t like it, you can just leave it.

 

But of course, if you just want to try it out on the free plan, you can try out everything on the free plan, essentially. So the only difference is really that on the paid plans, you get access to like a bigger, on the premium plan, you get access to like a better model and you get like more AI searches.

 

Um, uh, so you get like more AI credits that you can use and also more storage space as well.  So all of the data you own, we take zero ownership of the data. Uh, that’s also in the privacy agreement, in the terms of service that you can read. Okay. We don’t take any ownership of your data. The data belongs to you.

 

If you retain that ownership,  the same thing is true with the retrieval ability. You can retrieve all your data. It’s never stuck. So if you want to leave and use another AI tool, just download your data, use that tool. You can cancel your plan anytime if you don’t like it. You can just cancel anytime. All of the data is private, and then we use like industry standard security measures, of course, to keep it secure.

 

And we have customers from all over the world, from the US, where you guys are based.  Uh, we’re and of course here in the EU as well, we also have a lot of different customers from, uh, from, uh, Europe and Australia and Asia as well. So love to hear what you guys think. Um, here is my email. So if you want to reach out, uh, you can either actually let me write it in the chat.

 

Um, my personal email. You can use the contact form otherwise in the, uh, chat.  But I’ll, I’ll write this here. So you have adam@eno.com. That’s my personal email, so if you wanna get to me quickly, you can, you can just use that or you can just go to the contact form and somebody from our team will essentially help you out.

 

Um, okay. So with that being said, we can use, I can see if there are any questions. Uh, I’d love to hear your questions if you have any.  Uh, I’ll go from the top here. I came across, uh, I’m reading now from the comments.  So, uh, Recep said,  or is it Recep, maybe? Uh, maybe a Turkish name, correct me if I’m wrong. Um, I came across AI citing non existing articles, basically creating them and citing them.

 

Yeah, so this is the problem that I was talking about before. The newer models are much better. So, uh, we’re, like, in Avanote, we’re using whatever model we can find that is the best for giving academic citations. Um, in my view, I would say like the new GPT 4. 0 version is much, much better than the old versions.

 

But I think the best version right now for just the academic citations is probably the anthropic model. But even with those models, it will sometimes make up citations, like it will. That’s just a problem with AI that hasn’t been solved yet. I think it’s going to be just a matter of months until this has been solved.

 

But right now, when it comes to citations, you have to be extra critical. So the actual arguments will often be correct, but the citation will, a lot of the times or these days, it’s like 50 50 almost, will either be correct or it will be false. So you have to double check and make sure the citation is correct,  or if it’s not correct, you have to find like a real source to substantiate what was said.

 

Okay, Nicole said, amazing product. Thank you for the information. I have to go, but very interested in testing it out. So thank you so much, Nicole. If anybody else wants to try it out, you can just register on Avidote and just try it out on your own time. Uh, of course, if you want to have a trial, you can just email me and then we can, um, uh, get you a trial.

 

So that’s basically for departments and for research groups. Uh, if you don’t, if you’re not a research group leader, maybe you can talk to your department head and we can do a presentation for your team. And then, um, if your department wants, we of course could give discounts for the whole group. So the department basically, uh, covers the bill.

 

Uh, but of course if you’re if you don’t want the department to cover the bill, you can just buy it yourself Uh, we have discounts for students. Also. I should add that so if you’re a phd student, you can just email us and We have discounts for students.  Uh, do we have any list of specific syntax for asking questions?

 

How about the plagiarism check feature?  Uh, do we have any list of specific syntax for asking questions about plagiarism? Okay. So yeah, so we have specific Um,  what do you mean exactly? So I’m not sure I understood correctly. Do you have any list of specific syntax for asking questions? How about the plagiarism checkpicture?

 

So the plagiarism is already in the model itself is, is optimized in order to avoid plagiarizing. Um, like by, by default, if you give it like a proper input prompt by default, the visual, the response should be original, but there is a potential reality that one or two sentences could be copied from somewhere, which is why you always have to.

 

Uh, make sure that you incorporate your own text. So, from my perspective, you don’t really need to look at plagiarism, uh, checkers, because if you’re using AI correctly, you would never copy and just paste it and then use it that way. You would always rewrite it, you would always incorporate it within your own text.

 

And so if you’re using it in a proper way, you would never have to double check if that exists anywhere else, because you’ve already rewritten it in your own voice. So you wrote it, you know that it can’t possibly exist anywhere else.  Um,  so Roman said your presentation was great. Are there tutorials available for users?

 

Uh, what is the cost for paid plans? Uh, yeah, so we have a 19 per month plan or 200. So you save like two months if you pay the yearly plan. Uh, so if you pay yearly, it’s 200, uh, you save two months. If you pay per month, it’s like 19. And then there’s also more, uh, advanced plans. So if you want to have the most advanced plan, it calls cost, uh, 59 per month.

 

Uh, so yeah, but I can send, I can send information later on to everybody who wants to use it. So it depends really on your budget and how much you’re willing to use it for. The 19 plan is enough for most people.  But if you want the most, uh, the biggest plan, you can, you can get the most expensive plan.

 

Okay, Mariana, thank you so much for saying great presentation. Thanks. Thank you to you as well. Uh, I’ll, I’ll think, uh, we have three more minutes. I’ll stay, I’ll stay for those three minutes. Uh, anybody who wants to leave, I won’t keep you hostage. Uh, even though I’m Swedish, there’s no Stockholm syndrome going on here.

 

So if anybody wants to leave, uh, you are more than welcome to leave. Um, actually plan to visit the U. S. later this year. So if anybody wants a live presentation, if I’m anywhere near you, you can just let me know. And then we can try to make something happen there.  Okay, I think that’s all. I don’t see any more questions.

 

I’ll send the recording, by the way, to everybody after the event, and you can share that with your team or anybody you think it’s useful.  So thank you everybody for coming.  I’m really glad you took the time to listen in today.  And, uh, I hope, uh, if you want to join any of our upcoming events, we plan to have some different events on, on focusing on different topics.

 

Uh, you’re more than welcome to sign up to those events as well. Uh, and in the chat, it says, uh, Muhammad Ali, that’s a very cool name. Uh, cash is clay Muhammad Ali. Uh, so, uh, uh, very nice, great presentation. Thank you for, for coming you as well. And then fungs, uh, uh, fung fung said, thank you. Sorry if I messed up your name.

 

Um, yeah. So I’ll, yeah. Does the AI cite the newest articles or books? Uh, yeah, it, it, it cites like, uh, uh, recent books up until like a year ago. That’s the knowledge cutoff date, but we’re working on improving that. So it’s even more recent. Okay. Thank you, everybody.

 

Take care