Enter your feedback on an AI Search Engine for a chance to win one of two $50 UofT Bookstore gift cards! 

GenAI search engines are search engines that typically: 

• use generativeAI to translate the user's natural language query into a new search (eg. Boolean query or embedding-based search)
• searches large open-access databases of scientific papers and pre-prints (eg. SemanticScholar, OpenAlex, Arxiv.org, PubMed)
• selects results from search based on semantic/topic match algorithms
• incorporates the use of "Retrieval Augmented Generation" to generate summaries of results and provide references that are grounded in external data sources

There has been a proliferation of AI search engines since the release of ChatGPT in 2022. When considering whether to use or invest in an AI search engine, find the tool's About or FAQ section and think about: 

• where does the data come from? Most AI search engines search SemanticScholar which derives its content from more than 50 direct partnerships with publishers, data providers, and aggregators. It includes content crawled from the open web as well as trusted indexes like PubMed. This massive dataset can be amazing for discovering outliers and difficult to find information; however, because it also crawls the open web for scholarly publications, it is perhaps more vulnerable to including fraudulent publications. 

• what can you do in the free version that you cannot do without a paid subscription? 
• what features are important to your work? How does the tool integrate with your existing technologies used for doing literature reviews (eg. reference managers, notetaking software).
• how transparent is the tool about:
  • the dataset being searched
  • how your query is being translated into a search strategy
  • how results are selected 
  • how results are being summarized