Once upon a time, it was incredibly difficult to locate an appropriate academic article. It was like attempting to locate one specific needle within a massive, infinite amount of hay. The only available tool at our disposal was to enter random combinations of search phrases into the digital landscape in the hopes that somewhere within the larger scope of digital data, some PDF would eventually be sucked toward us. Paper searching had significantly less engagement as a process and was primarily a one-way, desperate attempt of being heard in a multiverse of infinite emptiness. To locate an interesting academic paper, I could search for “2020 machine learning models for climate change,” hold my breath, and then review hundreds of very loosely related results, including both obscure conference papers and tangential blog posts. That was the era of Boolean searching, exact phrases, and an overwhelming number of lost connections; the search itself was typically the biggest hindrance to discovery. However, a monumental shift has occurred. In the past decade, searching for and obtaining knowledge has been changing; it has gone from being a rigid, word-driven chore to one that is much more “human” in nature. This transition from a paper-based to an electronic-based searching method has been an ongoing conversation between technology and humanity—where technology is learning to use our language with us in order to provide better access to the enormous body of scholarly literature than we could have imagined before.
The Age of the Keyword Hunt
For many years, keywords were essential for locating research. In academic databases, the organization was primarily by metadata, including titles, authors, journals, and index terms. In order to be able to navigate this world, you had to think like an academic database. Your natural curiosity, “I want to see what has been published recently regarding using mindfulness to enhance concentration while taking courses online,” must be distilled to the bare essentials. You had to be able to cut out the vagueness of your thoughts into single, searchable chunks of information: “mindfulness AND focus AND online learning AND students.” Finding papers was dependent upon both vocabulary and predictive skills: did the author use the term “attention” or the term “concentration”? Did they use “e-learning” or “digital education”? The use of a synonym could make the difference between finding a wealth of information or reaching a dead end while researching.
The technique resulted in barriers that were not readily perceivable to the person using them. Those who were already immersed in a field’s specific language had the greatest access to information, which made it difficult for others to access that same information. For example, a graduate student may know the precise term for a technical concept, but an interdisciplinary researcher or a professional from a different field might have difficulty locating or understanding the same term. The search process that was used to locate relevant publications was primarily transactional; you received what you were seeking. However, as a result of this transactional process, a person doing a search almost always failed to discover something of significance that they did not know to ask for because neither the searching process nor the tools used during the search provided any form of “chance encounters” with publications that were adjacent to their own that were either coincidental or purely happenstance. The searching process was very efficient if considered in isolation; however, through its limitation, the searching process could not facilitate the development of broader, more creative connections between intellectual disciplines. While the searching tools produced high-quality outcomes, they placed upon people the obligation to perform all of the heavy lifting needed for translational activities.
The Natural Language Revolution
The conversation is shifting to a more sophisticated AI environment thanks to the emergence of advanced large language models and new neural search tools – bridging the gap between the way we think and the way we conduct searches. This is more than simply matching keywords better; it also includes understanding. Major search platforms now have an increased ability to understand search intent, context and nuances within the textual content of a complete sentence rather than only matching individual words. You can now type your search query into the system as you would ask a well-informed colleague: “What are the most cited papers from the past three years indicating that gut microbiota has an effect on depression?” The AI will now be able to process the meaning of the search, analyse the search concepts (gut microbiota, depression, citation impact and time frame) and deliver papers that reference the concept rather than simply keywords found within the text of the papers themselves.
In this new era, searching for papers has been fundamentally democratized. The barrier to entry to search for papers has been significantly reduced compared to before. (Previously you had to have a “secret” code to access the paper search system, now all you need to access the system is to be curious enough to search.) This way of searching by using natural language mimics how we build and share knowledge (i.e., by asking questions, having conversations, and trying out different ways to ask for something) and changes the search box from a command line to a dialog box. The technology that powers this new approach includes the use of transformer models which understand relationships between words semantically, therefore enabling us to search for the same body of literature using different key words. When searching for papers, we will experience both less of an archeological dig and more of a tour through the complete library of mankind’s knowledge.
From Search to Discovery and Synthesis
The evolution of research tools will continue to progress beyond the answer to a question and toward the next phase of active synthesis and intelligent discovery. Modern research tools have transformed from simple retrieval engines into intelligent research assistants. Now when you pose a natural language question, the response will not only include a list of references but will also provide a brief summary of existing literature along with the key issues being debated; it will also trace out the history of thought related to the question; and finally it may even provide suggestions regarding how to view the original question in light of another perspective (based on a different discipline). As such, research now represents a new base from which to create genuinely new knowledge.
Suppose that you have been asked to look at the ethics around generative AI. If you were to use a traditional method of searching for papers on the topic, you might end up with 10,000 papers and then have to manually check through each one of them to see if they have what you need. However, if you were to use an advanced, developed system, it could outline several different ethical frameworks (e.g., bias, copyright, labour displacement) and identify the top researchers within each framework and point to the classic papers that started the discussion about that framework. Furthermore, it could point out missing references within the research field or identify ways to connect information that you would not have thought of prior to the search and this capability turns researchers into synthesizers to a much greater degree than in the past. The mental effort associated with conducting a literature search (i.e., filtering through, filtering out, worrying about missing key topics) is now available for higher-level logic such as analysis and critique and developing new concepts. The program provides a framework to build an understanding on to create your own contributions to the knowledge base.
The Human in the Loop
One of the questions this powerful technology raises about the skill of the researcher is whether it causes researchers to become lazy. The answer can be found by changing the way skills are defined. Memorizing the exact jargon will be less important. However, researchers will need to develop the ability to ask insightful questions. Critical thinking has risen. An AI has access to virtually limitless amounts of information; however, a human will determine the credibility of the information provided, evaluate its importance, and use that information to create a coherent story. By evolving how papers are searched for and located, we have created a powerful partnership between machine and human. The AI can examination millions of articles in seconds to locate and identify relationships and similarities between articles; however, the human provides the judgment needed to validate the information found, the curiosity and creativity needed to transform that information into knowledge, and, finally, knowledge into Wisdom.
This changed relationship creates a stronger emphasis onEvaluating Sources also, when results are returned in a conversational manner,It is critical to click-through; verify the journal; verify the methodology; verify the author’s background so that the ease of access and the vigilance to evaluate results are consistent. The purpose of this evolution in how we search for papers is not to replace the researcher but to provide them with assistance; remove the time consuming friction from the process of locating information so that they can focus on enjoying the process of discovery and performing rigorous critical analysis.
When you go online to learn something new, you are doing much more than entering information into a computer. You are participating in a long progression of human communication. From the aggravating game of finding the right keywords to the smooth flow of speaking with someone, the way people searched for papers has changed drastically over the years. Paper searching is now much more intuitive, powerful, and reflective of the natural rhythms associated with curiosity and exploration. Searching for a paper has become more than just the technology to search in different ways; it is about forming better connections between questions and answers, seekers and the vast array of beautiful things available to learn about this world. There is still an infinite amount of information; however, there are now guides who understand our needs and can help us discover new information.
