Artificial Intelligence in Education

Dorothée Chareron

, French Teacher, MYP and DP

Recent research highlights declines in cognitive engagement and memory retention because of the overuse of AI by young people. This is also the case for adults who overuse AI in their personal and work lives. AI is a great technology, but we all need to learn how to use it properly to avoid negative consequences on our brains, jobs and personal lives.

As educators, our responsibility is to teach and advise our students on how to use AI effectively in their studies so that they are ready for tomorrow. We can only do so if we are up to date with the latest advancements of AI and how it can support us in our jobs.

I’ve recently completed an online course on AI in Education with King’s College London. This course covers the following topics:

• What’s generative AI and how it works;

• The limitations and challenges of AI for feedback, assessments, teaching and learning;

• How to design AI policies in educational institutions and what to take into account when doing it

• How AI is impacting the work landscape;

• The considerations of social and ethical use of AI at work and in education.

The limitations and challenges of AI for feedback, assessments, teaching and learning;

Feedback is important and needs to be personalised. Educators cannot expect AI to write it for them. However, it's possible to note ideas on what students did well and what they need to do next to progress. AI can then help organise the feedback into a nicely written paragraph without forgetting personal feedback and comments.

As for assessments, educators can use AI to support them to create questions which engage students' cognitive skills as much as possible and develop their critical thinking. Offering different choices of topics and fostering creativity will also help students use less AI and more thinking, ideally. For teachers, it’s an opportunity to ask AI to evaluate their assessments before testing to see what changes they need to make. AI makes a lot of mistakes and a first trial isn't always the best one.

Marking assessments with AI can be a bit risky unless it is specifically trained for detailed tasks. It’s possible to have an in-school AI which isn’t connected to the internet to protect students’ privacy and work. This takes time and dedication at the beginning and lots of errors will happen. It also implies hundreds of trials with hundreds of students’ work, uploaded with their permission. Educators cannot upload students’ work to an AI programme without their consent: copyright and intellectual properties shouldn’t be ignored. This is what we are teaching our students after all.

In order to integrate AI into teaching efficiently, this online course suggests using the PAIR (Problem AI Interaction Reflection) framework. This helps integrate AI better into the curriculum with examples of real world scenarios which can easily be created with AI. This framework also teaches students how to use AI ethically while exploring and experimenting new work techniques with AI. 

Creating a chatbot for students is also an idea to help them work independently. This implies that teachers will need to feed the bot with relevant information related to what students can ask to understand a concept better for example. Students’ questions will also train the AI programme. The bot can also be internal to the institution.

Overall AI should only be used as an assistance rather than a machine doing the work for teachers and students. Many things should be taken into consideration as well before sharing anything with an AI programme.

How to design AI policies in educational institutions and what to take into account when doing it.

All of the above is closely related to a school AI policy. Everything needs to be clear from the start about the dos and dont's. These questions can help:

• What are students/educators allowed and not allowed to do when using generative or assistive technologies. This refers to students’ work and when creating/correcting assessments

• To what extent do we consider that a student has used “too much AI” for a task. What is the school doing to support students to use AI more effectively and ethically

• What are the limitations of AI detectors: discussions with students are a must to understand what they are capable of by themselves

• Are all staff trained properly to support students with AI and are they able to use it effectively for work

• What AI programmes should and shouldn’t be used by students in a school. Some platforms are specifically designed for the younger learners protecting them from sensitive information

• If a school decides to have a chatbot available to all students, what information should be entered and how to monitor what students are writing in the bot.

 

How AI is impacting the work landscape

AI is already replacing people or jobs but employees and workers won’t be replaced if they are up to date with AI. New jobs are already created as well (AI trainers, ethics specialists or machine learning engineers are some examples). Job roles are also changing with AI and can enable employees to work faster on repetitive tasks for example.

AI is comparable to the industrial revolution which scared a lot of people and replaced jobs. The big difference between both revolutions is that AI progresses much faster than the industrial revolution, and this is what scares some people.

The considerations of social and ethical use of AI at work and in education.

AI is great but some people forget that posting private information online can be dangerous for people’s privacy and confidential data. Having an internal AI not connected to a worldwide network is a clever way to bypass this, as long as it’s protected properly. 

Like any other product that young students use, there must be an adult around to check what answers are given by AI and if it is age appropriate for student’s safety.

Let’s not forget as well that some schools and students can have an unfair advantage with AI when they can afford premium versions which could give better output or options. However, prompt input is also important and this has to be taught to all students. Very good results can be given on a free AI version with very precise prompts and a follow up discussion with AI. The same chances should be given to students, no matter what their social background is. 

AI can finally be used to support students who need extra help. Here again, if educators can guide students to input the appropriate prompts, even in a free AI version, the result can be very helpful. 

While I was aware of most of the course content, the parts on the impact of the work landscape and the consideration of social and ethical use of AI were eye-openers. The AI policy design section was also a useful guidance on how to create one. I therefore recommend this online course to any educators or people working in educational institutions.