What is (generative) AI?

Generative artificial intelligence (GenAI) and GenAI have become a topic of much discussion and attention since the introduction of ChatGPT in November of 2022. Two months after its launch it had already gained 100 million users, ensuring that it is the fastest growing application in history - faster even than TikTok (Hu, 2023).  

GenAI can take many forms, such as the generation of text, images, audio, video, or other significant output. This should not be confused with the more ‘analytical’ AI which can be used to, for example, detect cancer cells. The most notable aspects of GenAI is the ability to generate the output based on prompts (instructions given to the model) in natural language (languages such as English, Dutch, Chinese, rather than binary computer language).  

The launch of ChatGPT has sparked a discussion on what the strengths, weaknesses, limitations, and opportunities of generative AI tools are and how we should approach them. This is a discussion that is especially relevant in education and research. At WUR we expect GenAI is here to stay, but it is up to us as members of the academic community to put it to good use and do so mindfully, ethically, and responsibly. These pages provide you with information on some GenAI tools, the rules and guidelines set for PhD candidates, the Examining Boards’ rules regarding the use of AI in education, and our guidelines on responsible use for WUR employees in general. 

Artificial intelligence (AI): “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings” (Copeland, 2023) 

Generative artificial intelligence (GenAI): Subset of AI that have the ability to generate data (text, images, etc.), often through natural language conversational interfaces. 

What can I use GenAI for?

For students it is not allowed to use AI for generating ready-made content (see rules Examining Board), and for PhD candidates there are rules and regulations as well. But there are many ways in which AI-tools can be a helpful addition to your toolkit, while adhering to the rules. For example, you might use AI: 

  • as brainstorming tool. You could use AI to provide inspiration when coming up with creative ideas. 
  • to improve your own writing. White AI tools may not be used to write full texts or paragraphs, you can use them to improve your own writing by asking for personalized feedback on it. Rewriting of texts to improve readability or to bridge language barriers is a useful and promoted practice.
  • for practicing for debates/discussions. You can prepare for discussions or debates by having an AI tool act as your opponent. 

Be aware that AI-models do not have an implicit understanding of a subject and may make significant errors due to insufficient understanding of a topic.

  • in literature research. You can use literature-focused AI tools to orientate yourself on relevant academic literature on your chosen topic.  
  • to improve or debug your code. To language models, code is just another language, and as such they can help write and improve code, and to fix errors.
  • to generate images. You can create images as supporting images for presentations. Do not use these images to replace photographs you could have taken yourself, however. 
  • … and many more.

Using AI-generated images in scientific publications is prohibited pending legal challenges surrounding the potential infringement of copyright by the developers of these models.

Common terminology

When using GenAI there are a few terms that are commonly used. Knowing these may help you better understand the technology and the rules applied to them better. 

Prompt: A question or instruction given to a generative AI model, typically written in natural language (human-spoken languages such as English, Dutch or Chinese). 

Prompt engineering: The practice of precisely formulating your prompt to most effectively provide a task or question to an AI model. 

Token: A single word or punctuation mark.  

Context length: The maximum ‘memory’ of a model, expressed in the number of tokens. This memory includes both the input and output of the model. 

LLMs: Large Language Models, a collective term for modern textual generative AI models. 

Inference: The process where a trained AI model applies its learned knowledge to new situations to make predictions, decisions, or to generate content. 

Questions?

If you have any AI-related questions, concerns, or comments after reading this information, there are several people you can approach. For questions on AI in literature research, you can contact the library ([email protected]). For general questions, concerns or suggestions about the use of GenAI within the WUR, feel free to email us at [email protected]. We would like to hear what’s on your mind!

References

Copeland, B. (2023, July 3). Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts. Encyclopedia Britannica. https://www.britannica.com/technology/artificial-intelligence

Hu, K. (2023, February 2). ChatGPT sets record for fastest-growing user base - analyst note. Reuters. https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/