03 January 2024

Teaching about AI and Machine Learning at KS3

Written by

Victoria Temple

The first CAS AI Community event welcomed Jane Waite and Ben Garside to share their work on developing a new resource set, Experience AI and their valuable insights into understanding how to incorporate AI into education.

“We had a really good attendance from CAS members, who, like me, enjoyed their informative and engaging presentation,” she said. “It was a great way to launch a new CAS AI Community and is exactly what we’re about – sharing ideas, inspiration and innovation.”

- Becci Peters, leader of the new CAS AI Community

Jane Waite, senior research scientist at the Raspberry Pi Computing Education Centre and Ben Garside, and Learning Manager at Raspberry Pi Foundation, outlined the design and principles behind Experience AI.

The resources are an introductory unit of work for KS3 students designed to teach about AI and Machine Learning, aimed at 11 to 14-year-olds.

With 50,000 units already downloaded, the resources already have received positive feedback from pilot schools.

“We’ve aimed at this age range initially, as this is the stage where teachers have just a little bit more flexibility with the curriculum and can choose to focus on AI,” said Ben.

“It’s also a formative time for young people who are making choices about their futures about their GCSE choices.

“It is also important that we look at the role of AI in the world, improve awareness about AI-related careers and to understand the involvement of AI in any career going forward.

“All the units have been piloted in the UK and feedback from pilot schools has been crucial in shaping the resources. We provide everything a teacher with no technical background needs to deliver the lessons. AI is not the preserve of computer science. Our resources are set out so that they are accessible and deliverable by any teacher in any subject or environment.”

The Experience AI resource set is research-informed with all the resources developed between learning and research teams at Raspberry Pi Foundation in collaboration with Google DeepMind - industry leaders in the field.

Jane explained; “It's a real collaboration. We consider them to be the industry leaders in AI so it's been fantastic to be able to work with them for their input. They've been involved in the design process and some of the quality assurance.”

Teacher facing lesson intro video
Lesson plan
Slide deck
Activity worksheets
Student facing concept videos
Assessment

Clearly define the learning objectives and limit these to a small number per lesson / experience.
Use the SEAME model to frame the learning objectives and time allocation of learning activities.
Always be informed by research on subject knowledge and pedagogy.
Be intentional about the use of vocabulary and concepts.
Avoid anthropomorphising in text and images.
Be specific for the age-group being targeted.
Include assessment.
Assume lessons are taught by non-specialist teachers.
Provide teachers with real world examples.
Follow a semantic wave.
Choose contexts that are applicable to all gender, ethnic, and socioeconomic groups.
The AI learning levels framework (SEAME)

Find out more about Experience AI here

Teaching about AI and ML – Language matters

Jane explained the use of language to talk about AI and ML is particularly important for this new field and had been considered carefully in the development of the Experience AI resource set.

“We need to think carefully about the language of this subject as it’s a new subject to teach, it’s important that we’re using the same consistent language across our field. We found a real muddle in terms of language, relating to AI and ML.”

Jane explained that there is increasing evidence that using anthropomorphic language distracts from the understanding that “human beings are responsible for AI and that modellers build AI models to produce outcomes. AI models are just a tool we use that create output.”

“Anthropomorphising reduces students desire to take an active role and be involved. This is also an opportunity to address gender inequalities. Let’s not make the mistakes of the past and let’s treat AI for what it is, a tool.”

Avoid using language that:

Distracts from the understanding that it is people who design and influence the uses of AI applications.
Risks a reduction in the students’ desire to take an active role in wanting to understand how they work and be involved in designing future applications.
Perpetuates dangers of AI in terms of bias and inequality

Avoid

Using phrases such as “AI learns” or “AI/ML does”

Instead use

phrases such as “AI applications are designed to” or “AI developers build applications that…”

Avoid

Words that are used to describe the behaviour of people (seen, look, recognise, create, make)

Instead use

Use system type words (Detect, input, pattern match, generate, produce)

Avoid

Using AI/ML as a countable noun

Instead use

Refer to ‘AI/ML’ as a scientific discipline, similarly to how you might use the term “biology”.

Find out more and get involved!

Join the CAS AI Community for regular events and insights into the development of AI.

Details here.

Ben Garside will be leading a session on Experience AI resources at CAS Conference in July