Data protection and copyright
- consider data protection and copyright when using AI
- do not give the AI any personal information, confidential or sensitive material, as it may store this information and use it to train the AI. The use of the work version of M365 Copilot Chat is recommended at SEAMK because this version does not share the provided information further. Microsoft also promises that the material it processes is not used to train the language model. Nevertheless, it is always advisable to be cautious about what information is given to the AI.
- do not feed the AI material created by others without permission. Data from licensed databases acquired by the library, such as scholarly articles, must not be fed into language models or other AI tools. Please check the FinELib website or E-library for the user rights.
- do not present the material produced by the AI as your own
Checking responses
- do not copy the answers provided by the AI anywhere without checking them
- critically evaluate AI responses for bias and misinformation
Sustainable use of natural resources
- Avoid unnecessary searches.For example, the energy and water consumption of a ChatGPT query is significantly higher than that of a traditional Google search.
- Electricity usage: a single ChatGPT query uses about 10 times more electricity than a Google search.
- Water usage: minimal due to low cooling needs. ChatGPT consumes about 1.4 litres of water per query depending on the data center location and the cooling method.
- Environmental impact: Emissions from a ChatGPT query can be up to 200 times higher than a traditional Google search, especially when AI-generated summaries are involved.
Follow the instructions for use artificial intelligence at SEAMK and Arene's recommendations on the use of artificial intelligence or universities of applied sciences.
