Consensus: Home

Consensus is an academic search engine for which we have an enterprise license, providing all UVA users access. It uses artificial intelligence to surface research findings which are relevant to students' questions.

The search engine used by Consensus is built over the Semantic Scholar dataset (which includes 200 million peer-reviewed documents).

Note that the results from a Consensus search are not meant to be taken as a final truth. They instead represent a snapshot of relevant research findings related to the question posed by the student.

Consensus uses non-generative AI to surface relevant references to the research prompt. It also can use generative AI to summarize the results.

Consensus is intended to make the process of finding peer-reviewed information that can be included in papers or research projects more efficient.

Instead of delivering a list of links like a traditional search engine, Consensus instantly extracts contrasting or supporting evidence directly from academic papers with a single search.

By clicking on a finding, you will be directed to the abstract of a paper with an option to view the full text if available (as in a traditional search experience).

Visit Consensus at https://consensus.app

Take advantage of the UVA Enterprise subscription by creating an account with your virginia.edu email address.

When formulating a search query, consider these three query types in order to make the best use of Consensus:

• Direct research questions: "does the death penalty reduce crime rates?"
• Two concept relationship: "mindfulness mediation and anxiety"
• Open-ended phrases: "immigration effects on the economy"

While Consensus will function with basic keyword searches, it is in fact built to accept natural language research questions. You will see better performance if you used a fully-formed question.

You can find more searching best practices on the Consensus website: https://consensus.app/home/blog/maximize-your-consensus-experience-with-these-best-practices/

Consensus will return relevant results, with the paper title, journal name, and year listed. There is also a sharing link for each result.

Some results are tagged "Highly Cited" or "Very Highly Cited" with an explanation that while the paper is highly cited, it does not necessarily mean that the results are any more valid than those from other papers.

Another tag you may find is "Rigorous Journal" which means that the source journal is ranked in the top 50% of journals on the SciScore Rigor and Transparency Index (RTI), described here: https://www.sciscore.com/rti/overview.php  

Other tags include "Literature Review" "Systematic Review" and "Meta Analysis"

Consensus also displays a "Study snapshot" for each returned citation which tallies the presence (or absence) of 7 key study attributes: Methods, Outcomes, Population, Sample Size, Duration, Location, and Results.

The filters provided by Consensus allow you to narrow down the search results (by date of paper, by number of citations for a paper, or by the Methods of the paper, among other options.
 

You can also enable the "Synthesis" and "Copilot" features, which engage a bit more with AI to analyze and synthesize the results from the papers to try to give you an overall sense of the results. The full power of these features are only available on Open Access papers, which have a PDF that Consensus can analyze. Of course, you will want to check these results carefully, to be sure that the AI has drawn the correct conclusions from the papers, but this can be a helpful start. The Copilot feature (not related to Microsoft Copilot) allows you to pose questions, and ask Copilot to draft content, create lists, etc.