Can Google Scholar survive the AI ​​revolution?

Can Google Scholar survive the AI ​​revolution?

Google Scholar, the largest and most comprehensive scientific search engine, celebrates its 20th birthday this week. In the past two decades, some researchers have found that this tool has become one of the most important instruments in science. In recent years, however, competitors have emerged that uses artificial intelligence (AI) to improve the search experience, as well as others who enable users to download their data.

The effects of Google Scholar, which is operated by the internet giant Google in Mountain View, California, are remarkable, says Jevin West, a Computational Social Scientist at the University of Washington in Seattle, which uses the database daily. But "if there was ever a moment when Google Scholar could be replaced as a main search engine, this could now be the case, due to some of these new tools and the innovations that take place in other places," says West.

Many of the advantages of Google Scholar - the free access, the width of the information and the sophisticated search options - "are now shared by other platforms," ​​says Alberto Martín Martín, bibliometricist at the University of Granada in Spain.

AI-based chatbots such as chatt and other tools that use large language models have developed for some scientists into preferred applications when it comes to searching, checking and summarizing literature. Some researchers exchanged Google Scholar for these tools. "Until recently, Google Scholar was my standard search engine," says Aaron Tay, a scientific librarian at Singapore Management University. It still comes first on his list, but "I've started using other AI tools lately".

Nevertheless, given the size of Google Scholar and how deep it is anchored in the scientific community, "it would require a lot of effort to push it off the throne," added West.

Anurag Acharya , Co -founder of Google Scholar welcomes all efforts to find, understand and build on it more easily. "The more we can do, the better it is for the progress of science."

the largest and most comprehensive

Google Scholar step 2004 on the image area of ​​the literature search in appearance and changed everything . At that time, researchers used libraries to find information or were looking for academic papers via paid online services such as the citation database Web of Science. In the same month in which Google Scholar started, the paid service Scopus from Elsevier was also launched, an extensive database of scientific references and abstracts.

Google Scholar searched the web for scientific work of all kinds, such as book chapter, report, Preprints and web documents-including those in other languages ​​than English. The goal was to "make the researchers in the world more effective and to enable everyone to stand on a common border of science," says Acharya.

The agreements of Google Scholar with publishers give him incomparable access to the full texts of articles behind payment barriers - not only to titles and abstracts that offer most search engines. The articles are classified according to their relevance to a search query - as a rule, the most moved articles are brought to the top - and further search queries are proposed. The depth of the cover enables highly specific search.

Google did not announce any usage data for the service, but according to the web traffic meter Similarweb, Google Scholar receives over 100 million visits per month.

The database is also very good at pointing out users to free versions of an article, says Martín Martín. This promotes the open access movement, José Luis Ortega, a bibliometric at the Institute for Advanced Social Studies of the Spanish National Research Council in Córdoba.

However, Google Scholar is opaque in other ways. A central concern is the lack of insight as to which content, including which specialist journals, are searched and which algorithm is used to recommend items. It also limits mass downloads of its search results, which could be used for bibliometric analyzes. "We don't have much insight into one of the most valuable tools we have in science," says West.

Acharya explains that Google Scholar is mainly an addiction tool and its main goal is to help scientists find the most useful research.

updated search engines

Competitors have appeared in recent years that offer such bibliometric data, although none can outperform the size and access to full text items behind Google Scholar's payment barriers. A remarkable example is the Openalex, which started in 2022. In the previous year, the Microsoft Academic Graph, which searched the web according to scientific information, was discontinued and published its entire data set. Openalex builds scientific data on this and other open sources. Users can search the content that is cataloged according to authors, institutions and citations and also download the entire records free of charge. "You do what we had hoped for from Google Scholar," says Martín-Martín.

Another popular research tool, Semantic Scholar, was introduced in 2015 and uses KI to create readable summaries of work and identify the most relevant citations. Another tool, Consensus was launched in 2022 and uses the Semantic Scholar database to find answers to questions informed by research (West is a consultant for consensus). One of Tay Favorites is Undermind , which uses a sophisticated agent-based search, in which an autonomous entity scans scientific literature like a person and adapts the search based on the content found. It takes a few minutes - compared to seconds on Google Scholar - to deliver results, but Tay explains that the wait is worth it. "I think the quality of the results that come back is better than on Google Scholar."

Acharya says that Google Scholar also uses Ki to evaluate items, propose further search queries and recommend related articles. And at the beginning of this month, the company presented AI-generated article overviews for its PDF reader. Acharya adds that the addiction tool tries to understand the intention and the context behind an inquiry. This semantic search set is based on voice models and has been used for about two years, he says.

One thing that Google Scholar does not do yet is the inclusion of AI generated overviews to answers to a requested request, similar to those who can now be found at the top of a typical Google search. Acharya says that it is challenging to summarize conclusions from several works in a concise and context -rich way. "So far we have not seen a effective solution for this challenge," he adds.