Elicit is a research tool that helps users gather, analyze, and synthesize information efficiently, making it ideal for conducting literature reviews and organizing findings.
Elicit is an AI-powered research assistant developed by Ought that automates and streamlines the academic research workflow, enabling researchers to locate, analyze, and organize scholarly literature with unprecedented efficiency.[5] The platform searches over 138 million academic papers using semantic search and can analyze up to 20,000 data points simultaneously, offering capabilities including automated literature discovery, intelligent paper screening, systematic data extraction with sentence-level citations, and automatic generation of research reports inspired by systematic review methodologies.[1][3][6]
Elicit is particularly valuable for researchers in evidence-based fields such as biomedicine, social sciences, and education who need to synthesize large volumes of research, as well as for conducting formal systematic reviews and meta-analyses, with users reporting up to 80% time savings on systematic review workflows.[1][5] What distinguishes Elicit from general-purpose AI tools is its specialized focus on research accuracy and verifiability—it grounds all AI-generated claims in sentence-level citations from underlying sources, supports interactive multi-step workflows and large customizable tables that go beyond simple chat interfaces, and can draw from sources beyond academic publications including clinical trial data, regulatory documents, and product labels through its Research Agent feature.[1][2][4]
The platform accommodates researchers at all experience levels, from undergraduates uploading papers for expert review to senior researchers conducting large-scale meta-analyses, with collaborative features enabling real-time teamwork on complex research projects.[2][3]
The ToS states Elicit collects and aggregates user behavior data for service improvement and marketing, but explicitly l
No training opt-out mechanism was identified in the legal documents.
High risk across multiple dimensions. Only use with public, non-sensitive data.
The ToS states Elicit collects and aggregates user behavior data for service improvement and marketing, but explicitly limits this to non-identifying data. However, the scope of what constitutes 'identifying' is vague, and the DPA does not explicitly prohibit training on Customer Personal Data. The language suggests aggregated/anonymized data use rather than explicit training on raw user content, creating moderate ambiguity.
“[ToS] 'we collect and aggregate data concerning Service user behavior, traffic and other interactive and telemetric information...provided that such data does not identify you, and that you cannot be identified through further processing'”
The ToS grants Elicit a perpetual, irrevocable, worldwide, transferable license to use, display, perform, copy, modify, create derivative works, and distribute user content. This is an extremely broad license that survives account termination and permits sublicensing, creating high risk for company IP.
“[ToS] 'expressly grant us a limited, non-exclusive, worldwide, fully paid, perpetual, irrevocable, transferable right and license to use, display, perform, copy, modify, create derivative works, and distribute Your Content'”
The DPA requires Subprocessors to have data protection measures no less protective than the DPA and maintains a published Subprocessor List. However, the ToS references sharing with 'affiliates' and the DPA permits general authorization for Subprocessor use without explicit prior consent for each change, creating moderate concern about third-party access scope.
“[DPA] 'Elicit shall (i) enter into a written agreement with its Subprocessors, which includes data protection and security measures no less protective than the measures set forth in this DPA' and [DPA] 'Customer provides a general authorization to Elicit use of Subprocessors'”
The DPA states Elicit will retain Customer Personal Data for the term of the Agreement and any period after termination during which it processes such data, but does not specify a maximum retention period or automatic deletion timeline. The language is vague regarding post-termination retention duration.
“[DPA] 'Elicit will retain Customer Personal Data for the term of the Agreement and any period after the termination of expiry of the Agreement during which Elicit processes Customer Personal Data'”
The ToS explicitly grants Elicit perpetual, irrevocable, transferable rights to create derivative works from user content and distribute them. While the ToS states users retain ownership of content, the broad license granted to Elicit effectively allows the vendor to claim and sublicense derivative works, preventing exclusive company ownership.
“[ToS] 'perpetual, irrevocable, transferable right and license to use, display, perform, copy, modify, create derivative works, and distribute Your Content (including any derivative works)'”
The DPA references a Security Addendum at https://trust.elicit.com/ and mentions compliance with GDPR, UK GDPR, Swiss Data Protection Act, and CCPA. However, no specific SOC 2, ISO 27001, or HIPAA certifications are explicitly mentioned in the provided documents, and the security details are deferred to an external trust page.
“[DPA] 'Security Addendum means all additional controls and documents that support the protection of data, which can be found at https://trust.elicit.com/' and references to 'European Data Protection Laws' and 'CCPA'”
Track and measure your AI tools with Olakai's observability platform.
Explore Olakai