Like it or NOT: AI in Academic Research is here to STAY

..and it'll only grow more (powerful) from here... 💪 ....

Well, it’s a bit of a click-baity headline, but it’s absolutely true that academic research as we know it, is going to change drastically!

Hi to everyone in Gareth’s lovely community,

My name is Razia Aliani and I share my insights on the latest AI research tools and trends in my substack newsletter, ‘Era of the Research Bots 🤖’.

Today, as a guest author, I’ll be giving a compact, yet power-packed overview on the four steps in a Literature Review, and how and which AI tools are changing the game.

But before we dive in, here's a little secret: 

At the end of this email, there’s an exclusive gift 🎁 for you, the loyal followers of Gareth’s newsletter - something I haven’t even offered to my own followers yet 🙈

You wouldn't want to miss it!

AI has a bad reputation in the academic world. Don’t misuse it, instead lets work together to learn how to use it to your advantage. Here’s a workflow on how AI can enhance the literature review process across four key stages: Literature Search, Literature Mapping, Reading and Synthesizing, and Manuscript Writing.

I will take you through the process of how reviews are traditionally conducted and how AI can make you more efficient, productive and improve the quality of the research output.

1. Literature Search

Traditional Approach: The traditional approach involves manually searching databases, using keyword searches, and skimming through countless abstracts to identify relevant papers. This can be time-consuming and overwhelming.

AI-Enhanced Workflow:

1.1. Assess Opinion Diversity (Consensus, Scite) Start by using Consensus to understand the range of opinions and perspectives on your research topic. Use Scite to get the insights into how a particular research paper has been cited by other articles This helps in identifying key areas of interest.

1.2. Identify Key Concepts (Scispace, ChatGPT, Elicit) Use Scispace or Elicit to create a table of key concepts from the identified papers. ChatGPT can assist in refining these concepts further. Apply filters as necessary.

The above process can help narrow down the scope of your review

1.3. Visualize Concepts (ChatGPT, Show Me) Visualize the concepts using ChatGPT's "Show Me" GPT. This step helps in understanding how different concepts are linked in your review.

1.4. Set Alerts (R Discovery) Set up alerts to stay updated with the latest research without getting overwhelmed. R discovery will email you one most important paper per day in your research area. This ensures you don't miss any new developments in your field.

2. Literature Mapping

Traditional Approach: Literature mapping traditionally involves manually organizing and categorizing papers, often using spreadsheets or index cards.

AI-Enhanced Workflow:

2.1. Collect Relevant Literature (Research Rabbit) Use Research Rabbit to collect initial literature. Features like commenting and collaboration add a traditional touch, allowing for discussion and note-sharing.

2.2. Organize into Zotero Collections Organize your papers into collections. Research Rabbit allows for synchronization with Zotero, which helps in maintaining a well-organized database.

2.3. Assess Similarity Patterns (Litmaps) Use Litmaps for gap assessment and to identify impactful papers, authorship patterns, and keyword filters. This tool visualizes the citation network, helping you identify key papers and gaps in the literature.

2.4. Identify Insights (Inciteful) Use Inciteful for keyword filtering and similarity index. This helps in identifying crucial insights and connections between papers.

3. Reading and Synthesizing

Traditional Approach: Reading and synthesizing information involves manually reading through papers, taking notes, and summarizing findings.

AI-Enhanced Workflow:

3.1. Import PDFs and Create Folders (Scispace) Import all PDFs into Scispace and organize them into folders. Use the tool's co-pilot feature to search and ask questions from all PDFs.

3.2. Annotate Papers and Notetaking (Scispace) Annotate papers and save notes directly to your notebook. Don't neglect to read core papers thoroughly yourself; AI can miss subtle nuances.

3.3. Extract Data (Any LLM) Use any LLM to suggest the fields to extract data from, and to perform high-quality data extraction. ChatGPT is good with relevant GPTs, but Claude generally provides better results for this task.

3.4. Synthesize Data (Any LLM) Use Perplexity to synthesize data into meaningful themes, tables, summaries and insights.

4. Manuscript Writing

Traditional Approach: Writing a manuscript involves drafting, revising, editing and formatting the text. Researchers often struggle with writer’s block, structuring their manuscripts and ensuring consistency.

AI-Enhanced Workflow:

4.1. Formatting and Templates (Scispace) Select a journal before starting to write using tools like JANE. Use Scispace for selecting appropriate templates that match your target journal's requirements. This tool also helps in structuring your manuscript.

4.2. Outline Creation (Paperpal, LLMs) Begin drafting your manuscript using Paperpal's co-pilot feature. LLMs can assist in generating outlines and drafting specific sections. But they lack the depth so only use for overcoming writer’s block.

4.3. Paraphrasing, Proofreading, Grammar Check (Paperpal, LLMs) Use Paperpal to paraphrase sections of your manuscript. LLMs can help with academic tone and improving the structure. Use Paperpal for proofreading, grammar checks, and ensuring consistency in your manuscript.

4.4. Plagiarism Check and Submission Preparation Run plagiarism checks using tools like Turnitin or Paperpal. Prepare for submission by ensuring adherence to guidelines (submission checks in Paperpal can help), and drafting cover letters.

While all of this is a good overview on how to streamline your research workflow, this is just the tip of the iceberg. AI provides invaluable support for researchers and I understand that the sheer number of tools can be extremely overwhelming.

Which is why I have decided to do a live Masterclass on Research with AI on the coming weekend, where I will show you how to do MORE with LESS tools in this 3-day program. The masterclass is scheduled for the 24th, 25th & 26th May (see timings here) and 100+ researchers have already signed up!

Here’s where the special gift 🎁 comes in: drum rolls please

For the first 25 people who use the discount code GARETHDYKE25 at checkout, you’ll get an additional and exclusive 25% Off on the current fee of this program! 

And no, I wasn't lying when I said I haven't even offered this to my own followers yet!

FYI - I’ve been exploring AI in Research for 5+ years (yes, that’s even before Chatgpt; my thesis back in 2019 was on AI tools in Systematic Reviews). And you can check the glimpse of some related project reviews here.

Check out all the details on the Masterclass and Register here:

If you want to generally stay updated on how AI and AI tools are changing the landscape of academic research then follow my Newsletter or my community of 58k+ LinkedIn followers where I regularly share content on this topic.

Finally, while this is the my first post as a guest author here, Gareth and I will hopefully be collaborating more often to bring diverse value, knowledge and content for each of our communities. ✌️

Thank you Gareth, for giving me the opportunity to write here, and thank you everyone for reading patiently through this long email, phew!

Note from Gareth: You might be surprised to see a mention of a paid masterclass here, as I’m not usually a fan of paid workshops. However, I’ve known Razia for a while and can vouch for the quality of her content. She has put a lot of effort into creating her first masterclass, so I’m making an exception here. Just to be clear, this isn't a sponsored post and I’m not earning anything from this. I’m simply glad she’s offering a good discount to my followers. I highly recommend you register for this masterclass - there’s some really useful content being covered.