Google Scholar’s New AI Outline Tool Explained By Its Founder
Google Scholar has entered the AI revolution. Google Scholar PDF reader now utilizes generative AI powered by Google’s Gemini AI tool to create interactive outlines of research papers and provide direct links to sources within the paper. This is designed to make reading the relevant parts of the research paper more efficient, says Anurag Acharya, who co-founded Google Scholar on November 18, 2004, twenty years ago last month.
In honor of Google Scholar’s 20th anniversary, Acharya shares how teachers and their students can make best use of the new AI features available through the Chrome extension Google Scholar PDF reader.
A former professor of computer science at the University of California at Santa Barbara, Acharya grew up in India. As a student, he was frustrated by the lack of access to research materials available in India. When he came to the U.S., he says he stepped off the plane a better researcher than he was in India.
“I got access to resources, I didn’t become smarter,” he says.
But even in the U.S., access to much-needed scholarly material was often difficult to find for scholars and researchers in various fields. Acharya and his Google Scholar co-founder Alex Versta, realized that this was slowing down research. They decided to take the lessons they had learned by developing Google search and apply that to the world of academic papers, research, and study. The goal was to make it more efficient for all researchers to build upon existing research, Acharya says. It has also helped make academic work more accessible to students, educators, and general enthusiasts. But AI is now allowing the tool to take things one step further.
Utilizing AI For Deeper Research
“Scholar, for a long time, has been focused on helping you find things,” Acharya says. “There are lots of different ways in which we help people find things. Finding research is a key component, but reading and understanding and following up is another very significant part of building on other people’s research.”
The AI-powered Google Scholar PDF reader is designed to help people navigate each individual paper itself, Acharya says. It does this a number of ways.
For example, anyone who has done research has come across a citation they want to examine. But when you’re reading the paper this usually comes in the form of a bracketed reference, and you have to look it up in the citations section of the paper. Then you need to copy and paste that other paper’s name.
“You put it in some other search service and hopefully you can find some way to get to that paper,” Acharya says. Google Scholar PDF turns that initial reference into a link and makes navigating to that second paper easy.
Additionally, Google Scholar PDF uses AI to create an annotated table of contents for each paper. Typically, a table of contents is just section headings, says Acharya, but this creates quick descriptions of what’s in each section of the paper with bullet points. You can then click on these bullet points to navigate directly to that portion of the paper.
“So you can skim the parts that you want to skim, and you can go into detail for the parts you want to go into detail or decide that ‘I know enough about this and I don’t need it,’” Acharya says. He adds the tool might help make a long paper less intimidating for a student who is researching this type of topic for the first time.
The Future of Generative AI As An Aid To Research
Acharya says generative AI’s ability to understand language so well is one of the features that makes these models so powerful.
“The fundamental ability to understand language will enable us to do many more things,” Acharya says. “First is finding and reading. Second is to basically be able to comprehend an entire group.”
For Google Scholar, Acharya would like to see generative AI be able to quickly summarize research that is related to whatever paper a person happens to be reading. This could identify the research that seems to contradict that paper as well as new research that has appeared since it was published. Right now doing this isn’t possible, but Acharya has built his career on asking complicated and difficult questions and ultimately figuring out the answers to them.
Ultimately, Acharya is excited by the new era of generative AI. “It’s a wonderful time to be able to be participating in these efforts. There’s so much possible,” he says. “I think we have only scratched the surface, so I look forward to what is yet to come.”
Post Comment