Generative Artificial Intelligence and research tools are developing technologies. These rapid changes and the way they generative content can impact how they're used and any policies. Students should check with their professors and relevant university departments for guidelines.
The Leinhard School of Nursing provides a detailed artificial intelligence (AI) statement in the student handbook on page 28: https://www.pace.edu/sites/default/files/2025-07/chp-lsn-undergraduate-nursing-handbook.pdf
Types of AI Tools
We hear a lot about tools like Chat GPT, but not all artificial intelligence tools perform the same functions. It is important to be aware how different tools work in order to understand when they are useful or when they could fail.
| GPT (Generative Pretrained Transformer) | Agent | AI Powered Tool | Reasoner |
| A GPT is a language model trained to predict and generate human-like text based on input. | An agent is an AI helper that can use tools, look things up, and complete tasks for you based on a goal. | An AI-powered tool uses artificial intelligence to help with a specific task like editing or organizing information. | A reasoning tool is a version of GPT designed to solve complex problems or complete research step-by-step using logical or structured thinking. |
|
Tools: ChatGPT, Gemini, Claude Example Use: Prompting with, “Explain photosynthesis,” returns a well-written explanation of the process in text format. |
Tools: Amazon Recommendations, Google Maps Example Use: You tell an agent, “Help me register for classes that fit my schedule,” and it checks your school’s course catalog, avoids conflicts, and suggests a list of options. |
Tools: Grammarly, AskPDF, Canva Example Use: Grammarly’s targeted use is to check your essay for grammar and clarity, suggesting better word choices and fixing punctuation |
Tools: Deep Research, DeepSeek Example Use: Ask a reasoning GPT to “Write a report comparing the environmental and economic impacts of a city switching from gas-powered to electric buses.” |
Types of Search & AI Tools
"How are tools like LitMaps or SciSpace any different than using ChatGPT?"
In order to understand that difference, we need to explore the differences between types of searching:
| Lexical Searching | Semantic Searching | Neural Searching |
| There is a one-to-one match between your keywords and the articles it locates. | Search tool is able to consider the meaning and context of your keywords to locate relevant articles, so you can ask your query in the form of a question. | Uses neural networks (such as LLM transformer models) to understand the meaning of queries and documents, allowing it to retrieve results based on context and intent rather than exact keyword matching. |
|
Search = Depression AND Postpartum Results = Only returns articles that contain those keywords |
Search = "heart attack" Results = A search for “heart attack” will also return articles labeled with “myocardial infarction.” |
Search = What are the most effective treatments for postpartum depression? Results = Articles that address the potential answers to that question based on the context of the concepts |
So how do AI search tools differ from ChatGPT...
From the Types of AI chart, we can see that GPT tools are designed to generate human-like responses based on a question/prompt from the user. Tools like LitMaps or SciSpace are more like AI Powered Tools or Reasoners because they use AI-powered search to locate research articles and make connections between the metadata of the search results.
Example: SciSpace
"What does searching for literature using an AI powered tool look like?"
Below is a screenshot from the tool SciSpace which shows the following steps:
- Entering your research question in the search box at the top.
- SciSpace will returns a short summary of answers from the top 5 cited papers.
- Below the summary is a table of article results which displays citation details and brief insights from each paper.
- A column on the right allows you to add more columns of data from each paper (limited by your subscription type).
- Features such as "Request PDF", "Podcast", and "Chat" allow you to engage with the information quickly.
So how does AI enhance this searching experience and make it different from ChatGPT or Google?
- Your search results will use "neural search" meaning the entire context of your question is taken into consideration before producing search results
- Insights from the articles are extracted and put into an organized table format rather than having to reach each abstract
- The ability to quickly add Columns of different metadata points which will be instantly extracted and organized for each article
- Additional tools such as "Chat with PDF", podcast functions, and summary creations use AI to help you interact with the content
Concerns with AI Searching in Nursing Research
"So what role can literature review AI tools play in my research? What concerns are there?"
Literature review/Neural searching tools can be positive because:
- the neural search can help locate articles that may be missed during a traditional database search that uses only keyword/lexical searching
- the extraction of relevant information about each article can be time-saving
- some tools can aggregate information across articles and help locate gaps in the research around a topic
Literature review/Neural searching tools can be problematic because:
- the algorithms behind the tools are not fool-proof and any information provided should be fact-checked
- the data in the neural network that helps create the "context understanding" can be built on biased ideas which can contribute to bias in the search results that are returned
- the tools are searching a LIMITED number of databases and usually don't search pay-walled articles so you will not get a systematic, complete set of results
- they are providing CONTENT but not CONCLUSIONS. You as the researcher should be relying on your human research skills to create research conclusions.