The Writer’s Way
My Forced Marriage with AI: A Hesitant Writer’s Journey to Transformative Tools
A starter kit for tackling AI in writing, daily tasks, and career advancement
Discover how I, a hesitant writer, overcame my fear of AI and unlocked a new world of transformative writing tools. Since this is probably one of my most information-dense articles yet, here is a little overview of what this piece is going to cover:
- The marriage proposal — my story with AI
- The ring — The motivation behind tackling AI
- The optimal wedding planner — AI tools for writing, academic writing, and how I use them
- Learning the wedding dance — Tips and Tricks on how to get started with AI and apply the learnings to your life
The marriage proposal
I am not an eat-the-frog kind of person and neither do I like to follow the masses. I find myself most comfortable in that dark corner of the room where no one looks. But once I jump out of there saying something no one ever expected, I rejoice. I don’t like to take the obvious path, but I like to lure in the shadow and look for my opportunity to do something unexpected.
Therefore, when my professor asked me to write my thesis about a topic related to AI, because it is such a popular topic right now, I was disgraced. First of all, I know nuts and crackers about AI. Nuts in this allegory stand for “nothing” and crackers for “blown up sh***”.
Secondly, though I am sure I can come up with some weird idea on how to intertwine my actual interest with this made-up beast, I don’t think I can enjoy the process. Most likely, it will be A LOT of work. Extra work for learning something I don't know and extra emotional distress because I am not sure I have done my job right.
So why bother? Why force myself into a thesis-marriage with AI, when I can comfortably ignore it and write about something else?
The ring
Or so I thought when I hopped on a plane back to Austria. I introduced my thesis ideas to many of my friends, who are also studying at university. Notably, my home university is located in Taiwan, where technology and innovation are frequently evaluated and often integrated very quickly, at least in my opinion.
Austria is the flip side of that: Technology is viewed as the devil in the closet and AI its little baby. I came home to colleagues telling me they were not allowed to use AI. I thought this was funny and replied that they were bound to cheat then and use tools to hide their traces. Their reply was shocking, however:
“What do you mean do research with AI and AI tools for writing? My professor will catch me if I do!”
And that’s when it hit me: They knew so little about AI applications that had been openly discussed in my university, that they couldn’t even differentiate the different types of services out there. They were oblivious to the fact that writing a whole essay on ChatGPT was only one of many ways to apply the magic of AI in academic writing…
And this is when I felt like my marriage to AI might actually make sense. I realized that had I stayed in my old environment, I would’ve probably been oblivious still. A friend studying abroad in Australia shared a similar story with me and so I was convinced, that maybe just this once, I had to kiss the frog. It was time to prepare for my wedding to AI.
The optimal wedding planner
Now we jump to what I am using at the moment to leverage the learnings I have had so far. This is the part where you can get your fingers dirty copy-pasting useful AI tools for academia, because dang have I been busy on that front.
The god of AI tools in my opinion is Andy Stapleton. He posts regular YouTube videos on tools and ways to improve your academic writing with AI. Most of what I know and use on a regular basis stems from him.
My favorite AI tools in academia come in three categories: Search Engines, ChatBots, and Writing Tools. I will give you two examples for each, so you can get a better idea of what I am talking about.
Search Engines with AI are super powerful because with those you might sometimes catch an article or piece of literature, that would take you a long time to find on Google Scholar. Plus they enable more or less unpaid insights into the articles. These are my two favorites:
- SciSpace (https://typeset.io/): If I had to choose one favorite tool for academic research in this list, it would definitely be this website. There is little you cannot do, from chatting with documents to semantic search, snippets from articles and even building your own libraries of articles to compare. It can even write a small summary of your keywords at the start, with citations of articles that you can dig into.
- Research Rabbit (https://researchrabbitapp.com/home): This one is amazing because it is all free and offers you the opportunity to view articles as they connect together. You might be able to discover new keywords and generate networks between articles you read or discover other publications you were unaware of. Maybe the author used different wording but actually wrote about the same idea and now you can find the article and cite it. A moment of true joy for an academic!
As for ChatBots, I prefer to use ones that can accompany me on the journey through my browser or at least add information to what I can gather through downloading PDFs.
- Sider (https://sider.ai/): This one is my copilot for any task on the internet, which means I also use this a lot for non-academic writing. It is like a mini-chatGPT in my sidebar, accompanying me through my daily learnings. By drawing over a sentence with my curser I can ask it to read a section to me, improve it, keep writing, or brainstorm in the sidebar chat. I can also feed it some PDFs and have it work out a schedule for me, based on the website I am viewing and a PDF I read.
- PopAI (https://www.popai.pro/chatwithdoc): This one is more academic writing specific, but certainly has its applications outside of academics too. You might do an academic search here, or chat with your document in order to squeeze the information you want from it quickly. You might also want to create a presentation from a PDF or generate other output strictly related only to this one source. This is especially great if you don’t want the whole internet interfering with what the AI writes.
Finally, we get to writing tools, which is probably what everyone was excited about. To be honest, these are what I use the least for my actual writing unless my work has a maximum word count attached to it. I recommend that you use these tools for two purposes only: Getting into the flow of writing and perfecting an already-written piece.
If you use these to compile a complete text for publishing, you are DEFINITELY going to be flagged, and rightly so. It simply is completely unethical to write like this. However, I love that these can help me get that first sentence on a page or input citations when it is just too tedious to do it myself. Also, I suck at editing, so this is where all the cutting is done for me.
- Paperpal (https://edit.paperpal.com): This one works great for research, but I personally use it more to improve my fiction writing. Since I recently wrote a book, which required some cutting down of my wording, I have spent a lot of my time with this website. Shortening is really emotional for me, so I am often teary-eyed when using this service.
- JenniAI (https://app.jenni.ai/): This is arguably the OG of AI writing tools and if you haven’t seen an advert for it yet, you must be living on the moon. It not only provides writing support as stated above, but it very much writes for you, sentence by sentence. My favorite application for it is academic writing, specifically literature reviews. I need the flow that the AI gives me to formulate proper sentences while having A LOT of information from each paper in my mind. A true lifesaver for busy students, that has to be handled with caution.
Learning the wedding dance
And finally, there are some skills I still have to learn to have my wedding with AI go smoothly. This entails all the things that are currently on my “to learn” list and that should maybe end up on yours too.
Mainly these can be narrowed down to the following areas: Prompt engineering and an understanding of software engineering. I excluded coding in Python from this list because as someone who can barely use Scratch, I think I am millennia from attempting to learn that skill successfully. I am not saying it is not important, but at least it doesn't apply to me. You might still want to learn coding, though many now argue that this knowledge is becoming more and more obsolete through AI, so I will naively dismiss this gap.
For prompt engineering, there are many YouTube tutorials and Coursera classes available. It takes very little time to improve your prompting and the subsequent outcomes. But just knowing how to prompt only gets you half the way. Actually, this one simple video can already help you get a great kick-start to better prompts:
The main takeaway from this video is to include a persona, context, tasks, format, exemplar, and tone into your prompts, to generate more accurate outcomes. Once you have internalized the structure of this video and practiced it, you can watch the next video to extend your knowledge and prompt engineering abilities. Also, it helps to download the ChatGPT app, to speak your prompts, so it doesn't take you forever to type one out.
The tricky thing is, that you have to be fairly certain of the type of outcome you want if you are truly prompt engineering. As you will later discover, there is not one gold-standard prompt, but different tasks require different styles of prompts. A good prompt entails already knowing your target area well in advance, to pick the right mode of questioning. This is done in order to avoid “wrong” hallucinated output by the AI, which happens due to how large language models and complex neural networks work.
As for the second part, software engineering, I have to admit that I still have to learn my lesson. I have enrolled in a class for this semester that hopefully give me a good introduction because this field is hard to dig into as a complete newbie. If you want to embark on this journey yourself, you should google keywords like “large language models”, “deep learning”, “neural networks” and “machine learning”.
My advice for anyone trying to “get better at AI” is to really sign up for a beginner's course or an early introduction. This whole bubble around coding has become so intimidatingly huge, that as a nobody, you will have no chance to grasp it without a helping hand. Since AI is all over the place anyway, you will have no trouble finding such content on YouTube or Coursera.
Bonus Tip: If learning about AI itself is too big of a project at first and seems to not intersect with your life (which it definitely does trust me), you can also find an application of AI that would aid you in your daily life first and perfect that skill. This will eventually segue you into learning about AI on a grander scheme. For me, my starting point was to watch lots of videos on “how to use AI for language learning”. This is how I kickstarted my prompting journey and how I initially learned of the vastness of opportunities such tools have.
Post-engagement life
So now I am tied to this imaginary beast, with a ring and a plan and a deadline for our wedding. One year from now I will probably laugh about how little knowledge I had on all things AI. Most likely, many of you will also think that I am very naive. Trust me, I know. Many of you are probably far ahead of where I am, but that doesn't mean I shouldn’t be proud of embarking on this journey.
This is much too big of a project for me to tackle alone, and it is something I am utterly uncomfortable with. Still, I am writing this article not for those who already know, but for those who are also afraid. The ones who are sitting at their computers unaware of the magic that can unfold beyond ChatGPT. Students who are scared to google AI writing, because what if their supervisor found out? I want you to read this and think, that learning to put AI to work in your favor, is a manageable undertaking.
I am not portraying the image of me knowing about AI. Rather I would like to encourage others who are scared like me, to take this hurdle with me. There is no need to become a software engineer, but why not use the tools that are at your fingertips to improve your life a little? There is no harm in wanting to leverage the potential and maybe discover a new interest on the vast World Wide Web.
