Skip to content
Home » The Effortless Blog » Keep up with science using AI lit review tools and note-taking apps

Keep up with science using AI lit review tools and note-taking apps

I recently facilitated a session on integrating theory into practice as a researcher in industry. The participants’ main challenge was to keep up with academic developments. They share the same pain as the researchers in the academic setting. Both groups are:

  1. Feeling overwhelmed with tons of journal articles to read
  2. Drowning in information and failing to quickly identify what is relevant in their domain
  3. Clueless in selecting and prioritizing which journal articles to read
  4. Cannot find the information they need because their note-taking system has not changed since graduation
  5. Not up to date with the plethora of modern AI tools
  6. Cannot digest and understand the relevant information easily

There is a way out. We forget that a literature review is not about reading everything. Our role as scientists and researchers is to identify research gaps, ask critical questions, find answers, and contribute new insights. This is similar to the task of a detective and a literature review is just that, an investigation and not an article collection journey. We live in a time where AI literature tools and connected note-taking can dramatically reduce overwhelm and find relevant information that is easy to understand and fast. These are some steps I took to focus on critical analysis, generate insights using my voice, and become a literature review detective — not a summary machine or an article collector.


Join the Effortless Newsletter. Receive free tips on note-taking, literature review, AI tools, and other productivity topics specifically tailored for academics and students.


1. Leverage Chat-GPT to identify key questions in any domain

Thinking like a detective means that you have to be clear with the questions that you want to answer. Reading paralysis comes from doing it the reverse way — collecting articles and then asking questions later. This is a recipe for overwhelm. Be a ruthless detective. Be picky about your choices on the first go.

If you don’t know what your research questions are, put on your deer stalker hat and grab your magnifying lens to:

  • Clarify your research questions – write them down in your notes on a dedicated page
  • Check out open questions in your field – these are questions and problems that are known in the field but have yet to generate final solutions

Use these questions to search for articles or filter your article results.

If you need Dr. Watson as your sparring partner, Chat-GPT can be one.

For instance, in my social science project, some of the big questions I use as my reading guidelines include: what are the markers of an egalitarian society in the ancient past? How did people manage to govern without authority? When and how did we discard egalitarianism for autocratic modes of governance? I used this prompt to generate other parallel questions that I may not think of or that are relevant to the field:

Select which questions are relevant to you. From hundreds, I can whittle down my search and read a handful of papers that answer my specific questions. I based my prompt on a computational ecology example.

A small tip: You can specify how ChatGPT will answer your questions. To my taste, the default setting is overly verbose and too simple. Follow this short tutorial to set up custom instructions for ChatGPT.

No more collecting 100s of papers. Best of all, I’ve cut down on my reading time and I can focus on writing.

2. Identify relevant papers with special AI tools

Sometimes, all you have at hand is your first clue — one paper or one author. This is enough to embark on building a visual research and reading map using AI tools. Our brain responds best to visual maps to quickly comprehend information and get up to speed in any area. Some of the tools I have tried include Open Knowledge Maps, Connected Papers, Inciteful, Research Rabbit, Scite.ai, and Litmaps. Scite.ai is only available with a subscription while Litmaps and Connected Papers use a freemium pricing model. These new tools discover related articles even without using any keywords. For those in the medical or life sciences field, check out this post on how to use AI-based Evidence Hunt to scour through PubMed.

Here is an example of how Litmaps leverages a new way of displaying papers as a graph to help you discover relevant literature. Impactful papers appear on the right of the diagonal because they are recent (further right) and have many citations (further up). Having many citations in a short time is therefore often an indicator of impact or importance.

Litmaps graph showing importance vs relevance to identify papers

If you are starting and looking for clues in any domain, you can combine Google Scholar, Perplexity.ai, and Sci-space for instance. Your goal is to find a seed paper or a foundational piece of research that answers your research questions such as a meta-review paper or the latest research updates. Here’s a deep dive into finding seed papers on any topic. You can also ask questions directly on Consensus to receive answers from journal articles. For those embarking on new research and want a clean slate or a guide, consider the Effortless Literature Review course. This is a brief overview of the course.

3. Connect your thinking and ideas with note-taking apps

As a detective, your job is to piece together different forms of information as you begin reading, answering, and formulating more research questions. Your brain is getting excited and firing up all its idea pistons. How do you keep up with it and record new ideas as they arrive? I have ditched my old comfortable system of using linear formats like Microsoft Word and Google Docs. While they may have worked in the past, I was still losing ideas. I was forced to think in one way and that is vertical. I couldn’t branch out my ideas. All I got were multiple documents that I could not find because I only had document titles to rely on.

Our brain builds knowledge relationally and you need tools that allow you to document similarly. I’ve written about the advantages of existing tools in the market and how next-generation note-taking tools can dramatically improve your knowledge management. My choice is Obsidian. However, you can choose a wide array of connected note-taking apps such as Logseq and Roam Research for those who are comfortable in markdown formats. Others like Heptabase and Scrintal provide a whiteboard format for those who are used to the Miro experience. If you are used to Notion, Capacities might be something for you. Whatever you choose, the most important thing is to record every idea as you read and connect different ideas now or in the future. Build a robust knowledge management system that can grow with you as you learn more.

Expand your insights from one idea into infinite ideas

Sometimes we don’t know the relevance of information, yet. We need to turn all information into individual clues so that we can return to and follow them later. By creating atomic notes — or one idea per note— one data point becomes a clue that can be used over and over again to infinitely build concepts in the present and the future. This is the same technique that you already use, one Post-It or index card per idea. Except this time you will be doing it digitally and in a connected manner through your digital note-taking system.

Turn your one source into multiple atomic notes that build your conceptual understanding.

Minimise copying and pasting abstracts. Write in your own words. I found that my recall has improved and I can remember why the reading material was relevant once I did it that way. The result is a meaningful card catalog of your information, ideas, jottings, diary entries, meetings, and random notes that you can reuse when new research questions arise in multiple domains. Nothing will be forgotten or thrown out.

Learn how to write better notes using the Effortless Academic Manual system. Here’s a preview of the program.

Slash your reading time by curating an intent-to-collect list

Collect with purpose and compile a curated reading list. Make sure the reading list includes how you intend to use the material such as, ‘Crumley 1995 discusses the concept of heterarchy or hierarchy building and how to investigate it in archaeological data.’ Here’s an example of what an Intent to Collect reading list looks like on Obsidian.

Collect an "Intent to Collect" list with a notation to why you need to read this. It would help to filter out when the list expands.

Expect your reading list to expand. You can easily be overwhelmed. Be unafraid to delete entries on this list that are irrelevant or outside the scope of your problem or question. If you have not read the least important paper, chances are, you never will. Prioritize your time and effort. If you use Obsidian, here is a detailed recipe to filter your list.

4. Synthesize your knowledge using visual mapping

If you read crime novelist, Michael Connelly, and his police detective Hieronymus Bosch, you would have heard about the Los Angeles Police Detective Murder Book. This is every case’s bible and one of the important parts is the front page or exemplar page which provides an overview of what has happened in the case. A grand status update. This is similar to how you can put your clues and intuition together and create a research atlas.

Think of it as building a Wikipedia page of your knowledge so far. The next-generation note-taking apps make connecting ideas and notes much easier just like how Wikipedia does it. Every idea and concept possesses hyperlinks that open to another page or visually connect two cards to discuss a concept in detail. Fortunately, for those who are more visual, you can create visual graphs of your notes and ideas in any of the notetaking apps, or using Draw.io. I’ve also written about how to use your notes to visually identify inconsistencies and gaps. This is how a result might look:

An alternative to a Wikipedia style text summary, is a visual summary. This is an example using Draw.IO but you can also do this on Obsidian Canvas

By creating your master document this way you are linking your notes and putting together clues into a coherent whole. This makes formulating arguments, identifying conflicting opinions, missing data, or pursuing other research leads so much easier when everything is written in your own words already. Here is a method to leverage visual maps for creating arguments. You are building your muscle memory and knowledge by creating a growing outline of your investigation. Remember that research questions list we asked Chat-GPT earlier? You can use that to create your initial outline.

5. Develop confidence through free-writing

I cannot emphasize enough how free writing relieved my thinking from all types of fixed formats (especially journal structures). Think of this moment as your unedited voice coming out. This is why platforms such as Substack and Medium designed distraction-free writing platforms to help writers.

Write as you speak (…and speak well!). See how far you have gone and what questions remain unanswered after a round of reading.

  • Did you find answers to any of your previous questions?
  • What new questions arose?
  • What can you discard?
  • What gaps can you see?
  • Are you even asking the right questions?

Since initiating this change in my writing flow, I have noticed that I can structure my essays into my arguments and I converse better with the literature rather than parrot their ideas.

6. Automate your creativity with habits

Habits are the most difficult to break and develop. This is especially true when our old detection processes continue to work for us but deep inside we know it is not enough and we can do better, especially in the age of AI. We are traversing old and automated neural and behavioral pathways and we are establishing new ones. According to the study of Philippa Lally et al. (2010), habits form between 18 days to 8.4 months. It differs from person to person and from task to task.

Start small. Be patient.

The study also shows that any skips or falters do not affect overall habit formation. Repeat and refine your reading and writing workflow. Find your community and work together.

Summary

Modify this literature review workflow into your own and keep up with any domain without anxiety:

  1. Use ChatGPT to identify what interesting research questions exist in your field. Especially useful when you are new or interdisciplinary.
  2. Use AI tools like Scite, Litmaps, or SciSpace to find and filter through the plethora of papers released in your field.
  3. Note-Taking apps help you expand your knowledge horizon as you can outsource some of your memory into files and use a clever structure to find them again.
  4. Synthesize information visually instead of textually. This allows you to create an overview over much broader topics than with text alone.
  5. Develop a habit of free writing to bolster confidence and ability to express yourself as an academic.
  6. Stick to your path with the new tools. It will feel strange at first, but over a few weeks, it will become a second home to you and feel natural.

Want to get started on this journey?

I know that switching to new software or disrupting your current workflow is inconvenient and stressful. Start small. You can sign up for a free 8-day email course to learn the Obsidian basics. Moving forward the course on the AI literature review and the full course on effective note-taking is something to consider.