OpenEvidence is an AI-powered medical information platform designed to assist healthcare professionals, particularly physicians, by providing evidence-based answers to clinical questions at point of care. Created within the Mayo Clinic Platform Accelerate program, it aggregates and synthesizes medical literature from trusted sources such as PubMed, New England Journal of Medicine, and JAMA, covering over 35 million peer-reviewed publications in PubMed. OE offers features such as real-time literature searches, summarizing responses with citations, and tools for administrative tasks such as writing prior authorization letters.
Bottom line: For health sciences librarians, Open Evidence might support their work with health professionals. However, its underlying AI technologies raises concerns for those interested in scientific accuracy, transparency and rigour in performing point-of-care.
It is important to keep the distinction between searching for sources and searching for answers; Large Language Models provide the second while hiding the first!
Perplexity is a popular AI-powered chatbot (and search or "answer" engine) that provides AI-crawled answers with citations from "trusted" web sources. Perplexity has a deep research feature, and uses large language models (LLMs) such as GPT-4 and Claude to deliver concise, accurate responses. Unlike a traditional search engine, it focuses on natural language models, understanding text and providing predictable answers in a conversational tone. It’s available via web, iOS, Android, and browser extensions, with a Pro plan for advanced features. Its CEO is Aravind Srinivas.
Unlike conventional search engines, Perplexity employs a conversational interface, allowing users to ask questions and refine searches in a dialogue-like manner. The platform integrates web search to provide up-to-date information and includes citations to sources, enhancing transparency and credibility, though the exact selection of sources lacks clarity.
Bottom line: For health sciences librarians, Perplexity might support their work with health professionals but its underlying AI technologies raises concerns for those interested in scientific accuracy, transparency and rigour in performing reviews. It is also facing a number of lawsuits due to copyright infringement based on its underlying training data
Scite.ai is a literature discovery and evaluation platform that uses AI to find, analyze, and understand scientific publications. The platform requires signing up for a profile, which entitles you to three (3) free prompts. However, after it processes your prompt request or question, you have to provide a credit card number to see the extraction and report analysis.
The platform integrates Large language models (LLMs) into a research assistant that responds to natural language queries with cited evidence from literature sources. Its ”SmartCitations” feature extends traditional citation metrics by providing context regarding whether a paper supports, contrasts or mentions previous works. Scite.ai's corpus includes articles, book chapters, preprints, and datasets. Its AI-powered Research Assistant extracts structured data as PICO (Population, Intervention, Comparison,Outcome) terms or lists of biomarkers when set to Table Mode.
With the paid version, Scite.ai offers visual citation maps, customizable dashboards, real-time citation alerts, and a Reference Check feature that examines manuscript drafts for retracted or heavily contested sources. Its limitations are predictable: a lack of systematic search abilities, poor methodological reporting on how the articles were found, other than its citation retrieval capabilities. Retrieving non-peer-reviewed content limits its use in knowledge synthesis (KS) but may bolster grey searching.
These tools help researchers explore published data, summarize content, and answer questions based on uploaded documents.
These tools create images based on text descriptions, offering various options for customization and editing.
These tools generate presentations based on user inputs, allowing for customization of content, layout, and multimedia elements.
