Skip to Main Content
University of Maryland Libraries Logo
RESEARCH GUIDES

Artificial Intelligence (AI) and Scholarly Communications

This guide is intended to highlight the ways in which AI can be integrated into various aspects of the research and publishing process and what we should consider when deciding whether AI tools are the right fit for our individual practice

Artificial Intelligence and Scholarly Communications Graphic

Ethical Considerations

Graphic "Should I use Ai Tools?"

Ultimately, the decision about whether to use artificial intelligence in the course of your research and publishing activities is up to you. In addition to the legal landscape and publishers' policies, we can be guided by community consensus around the use of AI within our discipline. But since AI is a new and emerging technological space, that consensus may not exist. This portion of the guide highlights some of the ethical issues that we encourage you to consider in your decision-making, and additional guiding questions that can help to structure your deliberations or conversations with collaborators. 

 

+ Should I use AI tools?

  • Would you feel comfortable disclosing the tool and the manner in which it was employed?
  • Is there another tool or service that the AI replacing? 
    • Would you hire a copyeditor or designer to do this work or take it to a service provider on your campus?
    • Is this allowing you to contribute something new and valuable to your work - such as providing a transcript or other accessibility element?

 

+ Should I use AI-Generated materials?

  • Have you considered how much this output can be trusted? Can you trace where the information came from?
    • If citations are provided, are those accurate? Can you check them to verify?
    • If you’re using a research tool, what quality of databases is it pulling from?
  • Are you accounting for bias and inaccuracy?

 

+ Should I put my own or others’ work into AI tools?

  • Did you read the user agreement before signing up or utilizing the tool?  Do you trust the tools are you using and are you comfortable putting your work into them?
    • Assume what you put in is being used to train the machine learning model unless you are explicitly told it is not 
    • Make your own decisions with your creative and scholarly work but use extra caution and deference with others’ intellectual property

  • Could you create a dataset using public domain or openly licensed materials? 

  • Could you seek permission or license the content?

Labor and Environmental Concerns

Bar chart illustrating the difference between AI and human consumption of water and carbon dioxide emissions

 

Environmental impacts

Training a single AI model can use as much water as an average person uses over 27 years. Training and running AI models can produce carbon dioxide emissions that are more than 5x the amount a car emits over its lifetime.

Labor practices

Data labelling, transcription, and other human-performed tasks are generally outsourced and performed for low wages.

 

Resources