Artificial intelligence is becoming an integral part of our daily lives. Scientific communities around the world are debating whether it helps scientists in their work or, on the contrary, creates difficulties. In recent years, we have seen the active introduction of AI technologies on scientometrics database platforms. Today, we invite you to learn more about Scopus AI.
In today's world, people use artificial intelligence technologies in almost every aspect of human life, from everyday life and work to medical diagnostics. These technologies also play an important role for scientists, particularly in the preparation of scientific papers and in the course of their professional activities.
Artificial intelligence in Scopus
Scopus AI is a search tool that uses generative artificial intelligence (GenAI) to significantly improve the capabilities of the Scopus search engine. After a test launch in 2023, the full version became available in 2024.
Development and uniqueness of Scopus AI
Scopus AI is a special tool designed to provide more effective familiarisation with new topics, which, according to surveys, interests 60% of Scopus users. Thousands of researchers, librarians and academic experts from different countries participated in the development and testing process. Their feedback formed the basis for the creation of RAG Fusion technology, which significantly improves the accuracy of search results and the quality of response generation.
Scopus AI works exclusively with metadata and materials from the Scopus database published since 2003, allowing it to find the most relevant publications from the publishers represented in the system. The database is updated almost in real time, providing users with constant access to the latest and most relevant information.
Scopus also continues to work closely with the scientific community. For example, subscribers to the system are randomly invited to test existing and new tools. The development team carefully analyses the feedback received and responds quickly, often within a few hours.
Features of Scopus AI
| Search process optimisation | Scientists can obtain the necessary scientific materials in a matter of seconds in a systematic form that precisely matches their queries. |
| A reliable tool for selecting topics for future research | It can analyse large amounts of scientific content from the database in a matter of seconds and suggest relevant areas for new projects, as well as highlight areas of knowledge that are of particular interest to the scientific community. |
| Verified materials | Scopus AI provides access only to verified scientific materials indexed in its database. |
| Generation of answers | It creates research summaries and provides recommendations on the most influential works and leading experts in a particular field. |
| Concept maps | Scopus AI generates interactive concept maps that allow researchers to gain a comprehensive understanding of a scientific topic and its connections to other research areas. |
| Co-author search | The tool analyses over 21.9 million author profiles in Scopus to identify the best researchers in a particular scientific field. |
Using Scopus AI opens up many opportunities for authors. This tool increases the productivity and efficiency of research activities by providing authors with convenient tools for searching and analysing scientific content. Thanks to the use of artificial intelligence on the Scopus platform, scientists have quick access to up-to-date information, which improves the quality of their research and contributes to the overall development of science.
Types of content and processing algorithm in Scopus AI
Scopus AI only uses metadata, abstracts, and author profiles from the Scopus database to prepare responses. The final analysis takes into account the following content categories:
- Articles
- Books
- Book chapters
- Conference proceedings
- Reports
- Reviews
- Short surveys
- Data documents
Conference reviews and corrections are not taken into account. Additional measures were also taken to remove all articles that had been retracted. Scopus AI works with the entire Scopus database, selecting the optimal time frame depending on the specific query. For example, when searching for key works, the system analyses the entire array of publications to provide the most comprehensive overview of the main and influential studies on the selected topic.
Principles of responsible artificial intelligence at Scopus
- The real impact of decisions on people is taken into account.
- Measures are taken to prevent the creation or reinforcement of unfair biases.
- The principles behind decisions can be explained.
- Accountability is created through human oversight.
- Privacy is respected and robust data governance is upheld.
Scopus AI technology: transparency and reliability
The technology behind Scopus AI is designed with clear scope and understandable instructions. It ensures transparency, as all steps in optimising the query and obtaining relevant results are displayed on the screen in real time during the search, and the calculations used to identify new topics remain fully reproducible.
Reliability is guaranteed by the use of patented RAG Fusion technology in combination with the Copilot tool, which improves the quality of searches and responses. In addition, the large language model (LLM) operates within strict constraints set for prompt engineering, ensuring controlled and stable results.
Artificial intelligence provides researchers with new opportunities to increase research productivity and create high-quality scientific papers. At the same time, its use requires compliance with ethical standards, confidentiality, and responsibility for authorship. Although AI technologies are a powerful tool, they cannot replace the critical thinking and in-depth analysis necessary to achieve high-quality scientific results.
Combating various forms of bias
The issue of bias is taken very seriously at Scopus. Scopus AI uses only academic content from the Scopus database, which allows it to accurately identify the abstracts on which statements or conclusions are based. Algorithms select the documents that best match the query, ensuring that content is selected based on its ability to answer the question, rather than on the number of citations or the prestige of the journal.
If a query contains bias, there is a risk that it will be reflected in the response. Even in neutral queries, Scopus documents may contain some bias. To minimise this, Scopus AI is tested against two rigorous evaluation systems, including checking responses to questions on potentially biased topics to identify and reduce inappropriate results. Testing is performed on internal and external queries, for example based on Quora's classification of insincere questions.
An important role is played by a prompt system that filters out “dangerous” answers – those that may reinforce prejudice, harm, or stereotypes about certain individuals or groups. A feedback mechanism is also available, through which users can report harmful or biased answers. Reports are reviewed manually.
Want to publish a scientific article in Scopus? Forget about unnecessary worries – contact Scientific Publications! We will help you prepare your article for successful publication in the Scopus database and provide post-publication support. Have questions? Fill out the form below, and our manager will contact you shortly for a free consultation. Together to successful publication!
