AcademicID
AI Resources for Academia
AI Resource Library
A collection of open-source resources and links to helpful material for academics looking to utilise AI in their research.
We are continuously updating this collection. If you have any resources you would like to submit, please email us at [email protected]
In the rapidly evolving landscape of artificial intelligence (AI), this resource is for academics who want to integrate generative AI in their research responsibly and ethically. This document addresses the critical aspects of using generative AI in research, highlighting the importance of transparency, ethical considerations, and the maintenance of high academic standards.
The guide underscores that while the incorporation of generative AI in research processes, including hypothesis generation, data analysis, and dissemination, can significantly accelerate and enrich academic work, it is crucial to ensure the reproducibility and integrity of such research by being transparent about AI's involvement.
Key Guidelines for Using Generative AI in Academic Research:
- Transparency and Disclosure: Always disclose the use of generative AI, detailing its application, the models and tools used, and the prompts provided, to foster a clearer understanding of AI capabilities within the academic community.
- Human Oversight: Ensure human-led final decisions and maintain editorial control, emphasising the importance of fact-checking AI-generated content.
- Intellectual Property Respect: Be cautious of AI reproducing copyrighted material and ensure AI-generated content is original.
- Ethical Research Considerations: Adhere to standard research ethics, including the removal of personally identifying information and considering the societal impacts of AI usage.
- Reproducibility and Open Science: Aim for the use of open-source AI models, share methodologies, and ensure AI-generated content can be reproduced, enhancing the field's collective knowledge.
- Bias Mitigation: Be vigilant of inherent biases within AI models and take steps to mitigate potential biases to prevent discrimination.
Guidelines for Scientific Testing of AI Capabilities:
The document also outlines a framework for scientifically testing AI capabilities, advocating for comparisons to human baselines, robustness and reliability testing, and the interrogation of AI's decision-making processes. This approach aims to foster an open science environment, inviting scrutiny and debate within the research community while considering the social and ethical implications of AI technology.
Responsible Publication and Dissemination:
Lastly, the guidelines emphasise the importance of considering the dual-use of AI research, engaging in public dialogue to demystify AI capabilities, and continuously updating research practices in line with the latest advancements and ethical standards in AI development.
This guide is designed to serve as an invaluable resource for academics navigating the complex terrain of AI in research, promoting a responsible, ethical, and transparent approach to the integration of AI in academia.
The academicAI repository is a trove of tools leveraging generative AI to facilitate various academic research activities. From summarising documents to transcribing audio (such as interviews), these tools are built to streamline day-to-day research tasks. Each tool is accessible via a Jupyter notebook, ensuring ease of use and adaptability.
Current Tool List
New tools are regularly being introduced. If you have a tool you would like to have included, please feel free to open a PR in the repo or get in touch.
- Summarise documents: Automatically generate summaries and Excel spreadsheets detailing the contents of folders, including subdirectories.
- Page-specific summaries: Identify and summarise relevant pages from a collection of documents based on specified criteria.
- Transcribe audio: Convert audio or video files to text using OpenAI's Whisper Large model.
- Miscellaneous: A suite of general AI functions, including image generation and text-to-voice conversion.
Different journals have varying policies regarding the use of generative AI in academic research. Here are some key guidelines from prominent publishers and conferences:
- Springer Nature mandates that the use of Large Language Models (LLMs) should be documented in the Methods section. Outside a few exceptions, AI-generated images are not excepted.
- APA Journals' policy on generative AI requires that the use of Large Language Models (LLM) in drafting manuscripts must be disclosed in the methods section and cited. The full AI output must be uploaded as supplemental material.
- Science journals require the use of AI-assisted technologies to be disclosed in the cover letter and in the acknowledgments section of the manuscript with detailed information provided in the methods section.
- JAMA and the JAMA Network journals specify that the use of Generative AI for content creation or assistance in the writing/editing process is permissible, provided it is disclosed in the manuscript in a manner consistent with their requirements.
Below is a collection of additional resources for use of GenAI in academia.
- Michigan Institute for Data Science Generate AI Resources: A collection of great resources and papers on using generative AI in research. Includes example papers where generative AI use has been incorporated into research.