This is article number two, which deals with ethical considerations regarding using artificial intelligence in educational institutions. If you haven't read the first article, it's about the right to privacy.
This article will focus on the possibility of using ChatGPT as a knowledge database in different contexts at the school. The first part is about opportunities for the individual school to create artificial intelligence and what this can mean for how we access information. The second part of the article focuses on what happens when the major publishers start training artificial intelligence in their books. Here, completely new teaching aids will appear across disciplines and faculties, but is this what we want? In conclusion, there are some questions you can use to discuss the topic. We are not all supposed to agree on bans or an all-in strategy, but we need to discuss the options and issues before making a choice.
ChatGPT as the new intranet
Microsoft is working to open ChatGPT so educational institutions (and companies) can take advantage of the chatbot's many possibilities. This makes it possible to train the school's artificial intelligence with data otherwise located on many different platforms.
The dataset includes student data, meeting minutes, news, personnel data, guidelines, salary information, contact information, procedures, annual plans, etc. All the data we use today to run the school is otherwise reserved for the school's management to process and make available to the school's students and staff. This approach to contact with data may result in a superfluousness of the school's intranet as we know it today because management, staff, and students can ask questions directly to the chatbot, and then a qualified answer is formed. It will feel like you are conversing with a skilled and omniscient secretary, and there will be no restrictions on how many people have access to the resource simultaneously.
At the same time, once all this school data has been processed, it can easily be used in other contexts, and it will be possible, for example, to provide access to the Ministry of Education itself to extract relevant information from the school's database. Currently, many hours and money are spent on bureaucratic processes. A simple command to the school's chatbot will be able to provide a faster and more accurate response.
Much of the internal communication can thus be replaced by artificial intelligence, and it can help ease everyday life in schools. It can change the way we think about school administration. At the same time, it will affect communication at the school to such an extent that we can free up working hours for what is more interesting, namely teaching.
However, we need to think carefully before jumping on Microsoft's new chatbot because we move data into a closed dataset, where it may not be easy to get it out again. Schools can become tied up in expensive licenses and depend on a vendor and the services offered.
There is also a challenge in that we are not fully aware of how ChatGPT will answer questions and how rights management can be implemented in the system. On the web, there are quite a few examples of people getting it to answer something that should not be possible based on the rules that Microsoft has set. We will see the same challenge in schools because students will try all possibilities to test safety.
It will be difficult to secure data because although Microsoft may have a solution for this part, we will have to log all conversations. This, in turn, can cause challenges concerning monitoring employees. For example, an employee can ask the new chatbot about maternity rules; at the same time, management can follow these requests. You will also have to lay down rules about what you can ask them and how you ask.
We are very close to the above becoming a reality, and it can provide operational savings, but it also provides a distanced and impersonal approach throughout the education system.
Artificial intelligence as a learning tool
Another example is the large publisher's approach to artificial intelligence. Gyldendal, Praxis, and Systime can very easily train the chatbot with all their books and thus give school students an experience at ChatGPT. Thus, teachers and students can ask about the academic knowledge in the books and find content across subjects and faculties - while extracting additional information from the internet.
But what does it mean for our teaching materials that they are administered and perhaps moderated by artificial intelligence built on American values and rules? Should young boys and girls not have access to graphic content in sex education? Should there be blocks for searches of the Muhammad cartoons? Are Tove Ditlevsen's texts neglected in favor of Thomas Korsgaard? Will knowledge about the Holocaust only be available if the student is over the age of 18? We already see these challenges when schools use Google for Education because YouTube searches will filter out many videos, something we are not used to in a Danish context.
Conversely, it would also be great to have an artificial intelligence that knows the teaching material and can thus make quizzes, explain texts, set up connections across academic content, and, at the same time, help students by differentiating the content at different levels. There are many opportunities to make the material interesting and adapt the content to the individual student. However, we must consider what is important to learn as a human being and what we should leave to artificial intelligence.
The above are just examples of how artificial intelligence may affect educational institutions, and it is not a question of whether they will do so but more of when. Therefore, we must also decide at what level we want to use the technology in educational institutions because otherwise, the tech giants will sneak artificial intelligence into schools outside our control. We now have a choice, and some schools may go to extremes, where everything must be driven by artificial intelligence, while others will most likely remove some IT from teaching.
Question
- What data do we want to share with a company such as Microsoft?
- When do we become too dependent on the tech giants?
- How much do we want artificial intelligence to occupy space in our school system?
- When should we replace humans with chatbots?
- What happens when we break down our faculties and disciplines?
- How do we ensure Danish freedom of expression in American systems?
- How much should we monitor students and staff in educational institutions?
- To what extent will we let artificial intelligence take over teaching planning and teaching differentiation?
- Will we allow automated data pulling from school IT systems to save on IT staff?
- How do we ensure that students' data is not misused?