Elicit is an artificial intelligence (AI)–powered research assistant developed to facilitate and accelerate evidence-based inquiry. Its central purpose is to support researchers in conducting literature reviews, structuring research questions, and synthesising findings from academic studies without requiring extensive manual searching or coding. Developed by the non-profit research organisation Ought, Elicit uses large language models (LLMs) to automate the more repetitive and procedural aspects of research, such as identifying relevant papers, extracting key data, and summarising findings, while leaving interpretation and critical reasoning to the user.
In contrast to traditional search engines that return lists of papers based on keyword matching, Elicit is designed to understand the intent behind a user’s query. It draws from indexed academic databases to generate structured tables of evidence, highlighting how each paper contributes to the user’s research question. The overarching aim is not to replace human researchers but to augment human reasoning, thereby allowing scholars to dedicate more time to higher-order cognitive tasks such as analysis, synthesis, and argument development.
Elicit was first introduced in 2021 as a public beta, reflecting the culmination of Ought’s multi-year research into applying machine learning to scientific reasoning and decision-making. Founded in 2018 in San Francisco, Ought’s mission has centred on developing tools that make reasoning more transparent and efficient. Elicit quickly attracted attention in the academic and research-technology communities for its innovative approach to AI-assisted literature review and its transparent alignment with open science principles.
By 2022, Elicit had integrated more refined natural language processing (NLP) capabilities, improved data extraction pipelines, and expanded its connection to open-access databases such as Semantic Scholar. Through successive updates in 2023 and 2024, the platform incorporated larger datasets, improved prompt-engineering techniques, and launched advanced “workflow” features that automate parts of the research pipeline, marking its transition from a simple query engine to a full-fledged AI research environment.
Elicit offers a robust range of features that reflect both the needs of modern researchers and the ethical imperatives of responsible AI use. Key features include:
Literature Review Automation
Elicit assists researchers in discovering relevant papers by interpreting research questions posed in natural language. For example, a query such as “What are the effects of mindfulness-based therapy on depression?” prompts Elicit to retrieve peer-reviewed studies, display their metadata (title, year, journal, and authors), and summarise findings in a structured evidence table.
Data Extraction and Synthesis
The platform automatically extracts core elements from research papers, such as methods, sample size, intervention type, and key outcomes, and presents them in tabular form. This supports meta-analytic comparisons and systematic review preparation.
Semantic Search and Clustering
Elicit’s underlying AI models employ semantic rather than lexical search, meaning the system identifies conceptually relevant papers even when different terminologies are used. It groups papers by thematic similarity, providing an overview of the research landscape.
Question Decomposition
Beyond retrieval, Elicit helps users break complex research questions into smaller, answerable subcomponents. This feature aligns with the principles of systematic reasoning, allowing users to structure their inquiry logically.
Summarisation and Note Generation
The tool summarises key arguments and findings from each paper in plain language, assisting users who need quick overviews before deciding whether to read the full text. Summaries include reported limitations, helping researchers assess study reliability.
Integration and Export
Users can export tables, summaries, and bibliographic data into CSV or spreadsheet formats for use in research notes or citation managers. Integration with tools like Zotero and Google Sheets facilitates workflow continuity.
Collaborative Workflows
Elicit supports shared projects and allows team members to annotate or filter results, making it suitable for research groups, classrooms, and institutional reviews.
Overall, these features combine to make Elicit a dynamic bridge between automated information retrieval and human interpretation, an “intelligent assistant” rather than a “black box” search engine.
For students at undergraduate and postgraduate levels, Elicit provides a powerful, user-friendly means of conducting preliminary literature reviews and refining research questions. Its usability can be discussed in three principal dimensions:
Efficiency and Focus
Students often struggle to identify relevant papers among thousands of search results. Elicit’s structured interface displays summarised evidence in a tabular form, allowing users to focus on key contributions quickly. This fosters deeper engagement with fewer but more relevant papers.