Generative AI: Home

It is important to know how the use of Generative AI within education is seen within Zuyd:

• GenAI is not forbidden.
• Zuyd advises the use of Copilot (Microsoft 365) instead of ChatGPT. Log in with your Zuyd address
• Use other AI tools if desired only with a private email address.
• Students are always responsible for their own products. You choose whether and how to apply AI tools.
• You obviously take into account copyright and ensure correct source citations. 
• Be very cautious about using (privacy) sensitive information in GenAI. It is unclear what happens to this and where this information is stored.

Generative AI can have many advantages. GenAI does not necessarily do things better, but it does do them faster and more efficiently. One can think of a lot of applications where GenAI can be a useful tool:

• Answering questions
• Providing information
• Generating images
• Provide language and grammar assistance
• Rewriting texts to a specific target audience
• Creating translations
• Planning group tasks
• As a brainstorming partner
• ...

A critical attitude is required at all times. When do you make use of GenAI? What information are you sharing with the machine? What do you take for granted? In the ability to critically reflect on GenAI lies the key to ‘responsible’ use.

However, GenAI also has several risks. It is up to individual users to deal with these properly. Some of these risks, with options for dealing with them responsibly, are:

• Handling personal data (video): Sharing personal data via GenAI is undesirable. Be aware of what information you share.

► Make sure you anonymize your input.

• Respecting copyright and intellectual property (video): GenAI enriches itself with knowledge you ‘feed’. Be aware of this when using GenAI.

► Do not share information just like that. If necessary, ask permission to use someone else's texts as input or paraphrase (write it down in your own words). 

• Correctness of information (video): GenAI can make the biggest nonsense look like reliable information. GenAI writes texts that are likely and appropriate to the question/prompt. This does not necessarily mean that these answers are true. Moreover, because GenAI has knowledge of your previous questions, the answer may be personalized to the point where you end up in a filter bubble. 

► Always be critical and check if other sources confirm the information.

• Confirmation of stereotypes and discrimination (video): The vast majority of scientific literature comes from the Western world and, as a result, knowledge from the West is over-represented in the language models behind GenAI. This can lead to confirmation of stereotypes and even sexist or racist statements.

► Always read critically whether AI-generated texts are ‘neutral’ and correct them yourself or ask additional questions to fix discriminatory passages.