Skip to Main Content

Biological Sciences

This guide highlights resources relevant to research in Biological Sciences

Ask a Librarian

chat loading...

Introduction to AI

Generative AI tools have been receiving a lot of attention lately because they can create content like text, images, and music. These tools employ machine learning algorithms that can produce unique and sometimes unexpected results. Generative AI has opened up exciting possibilities in different fields, such as language models like GPT and image generators.

However, students need to approach these tools with awareness and responsibility. Here are some key points to consider:

  1. Information Hallucinations and Summarization Accuracy: Generative AI tools can produce inaccurate citations and summaries. Some tools use trusted sources, but users must understand the scope of these sources and verify all information independently.
  2. Novelty and Creativity: Generative AI tools can produce content that is both innovative and unexpected. They allow users to explore new ideas, generate unique artworks, and even compose original music. This novelty is one of their most exciting aspects. It also means that generative AI content is not reproducible.
  3. Ethical Considerations: While generative AI offers creative potential, it also raises ethical questions. Students should be aware of potential biases, sensitive data input, and the impact of their generated content. Responsible use involves considering the broader implications.
  4. Academic Integrity: When using generative AI tools for academic purposes, students should always consult their instructors and/or supervisors first. Always seek guidance to ensure compliance with academic integrity standards (please refer to the Duke Community Standard).
Disclaimer: The tools mentioned on this page are presented for informational purposes only and do not constitute an endorsement by Duke University. Do not use your Duke data, such as e-mail address, when creating an account as these tools are not licensed by the University nor have their data privacy policies and procedures been reviewed by Duke OIT. Users should approach the utility and application of these tools with caution and careful consideration. For more information, see Duke Information Security’s Use of AI Tools memo.

Research Rabbit

What is it?

Research Rabbit is a literature mapping tool that takes one paper and performs backward- and forward citation searching in addition to recommending "similar work." It scans the Web for publicly available content to build its "database" of work.

Best suited for...

Disciplines whose literature is primarily published in academic journals.

Considerations

  • Free!
  • Integrates with Zotero
  • Works mostly with just journal articles
  • Potential for bias in citation searching/mapping

» researchrabbit.ai «

Elicit

center

What is it?

Elicit is a tool that semi-automates time-intensive research processes, such as summarizing papers, extracting data, and synthesizing information. Elicit pulls academic literature from Semantic Scholar, an academic search engine that also uses machine learning to summarize information.

Best suited for...

Empirical research (i.g., the sciences, especially biomedicine).

Considerations

  • Both free and paid versions
  • Doesn't work well in identifying facts or in theoretical/non-empirical research (e.g., the humanities)
  • Potential biases in the natural language processing (NLP) algorithms
  • Summarized information and extracted data will still need to be critically analyzed and verified for accuracy by the user

» elicit.com «

Consensus

What is it?

Think of Consensus as ChatGPT for research! Consensus is "an AI-powered search engine designed to take in research questions, find relevant insights within research papers, and synthesize the results using the power of large language models" (Consensus.app).  Consensus runs its language model over its entire body of scientific literature (which is sourced from Semantic Scholar) and extracts the “key takeaway” from every paper.

Best suited for...

The social sciences and sciences (non-theoretical disciplines).

Considerations

  • Free and paid versions
  • Similar to Elicit, Consensus should not be used to ask questions about basic facts
  • Consensus recommends that you ask questions related to research that has already been conducted by scientists
  • Potential for biases in the input data from participants

» consensus.app «

Copilot (powered by GPT4)

Did you know that as Duke faculty, staff, and students, we have free access to GPT4 via Microsoft Copilot?

Log in with your Duke credentials to start using it today.


What is it?

Microsoft Copilot harnesses the power of GPT-4, one of the most robust large language models (LLMs), in the form of a chatbot, answering questions and generating text that sounds like it was written by a human. While not a replacement for conducting research, it can be helpful when it comes to brainstorming topics or research questions and also as a writing tool (rewriting or paraphrasing content, assessing tone, etc.).

Best suited for...

All users across all disciplines.

Considerations

  • Free way to use the GPT-4 model since ChatGPT-4 requires a subscription
  • Since it can't always verify its sources, be wary of hallucinations (or made-up citations) that can look very real.
  • It is not 100% accurate! While GPT-4 is touted as being 40% more accurate than its predecessor, users are still expected to verify the information generated by it.
  • There is always the potential for bias since the GPT model was trained on a massive dataset of websites, articles, books, etc. (much of which is inherently biased since it was created by humans).

Microsoft Copilot (GPT-4) » copilot.microsoft.com «

For ChatGPT-3.5 (free) » chat.openai.com «