Why teach AI for Good?

Artificial intelligence (AI) is often in the news. The public has increasing access to free machine learning (ML) and generative AI (GenAI) apps, so many people, including journalists and social media influencers, have been experimenting with them. The speed of change in GenAI has created new possibilities and creative outputs – AI appears rich with possibilities. Importantly, discussions on how the use of AI will improve efficiencies in society and the workplace raises questions about new job opportunities as well as creating a skills gap in the workforce. Even though everyday technologies have incorporated AI for some time, we have introduced an AI for Good course because there is now a general consensus on the value of teaching AI.

Ben Garside in How anthropomorphism hinders AI education explores why we should all avoid describing AI as human-like:

Not surprisingly, Apps for Good is all about doing good with technology – we believe in empowering young people to use technology to solve problems that they care about. There is the hope that, by harnessing the power of AI, humans will have more time to do good. Through our courses we encourage students to innovate, giving them the ideation tools to come up with original prototype app ideas. But we want young people to innovate responsibly and we recognise the emergent need for young people to develop their AI literacy – understanding AI so that they can make use of the opportunities, be mindful of the challenges and make well-informed choices about the use of AI.

In establishing the ‘SEAME Framework’ which offers a simple way to think about AI education resources to carry out a review of existing resources, Dr Jayne Waite summarises in Hello World’s Teaching & AI: "There are very few sets of lessons or units of work that were tailored to a specific age group. Activities often had vague learning objectives, or none at all. We rarely found associated assessment activities". Let’s explore how Apps for Good has addressed these identified needs.

As with other Apps for Good core courses, AI for Good is tailored to Key Stage 3 and S2 students, with detail on the key areas of: traditional programming to AI and ML, training data and GenAI, human-based bias and bias in AI as well as examples of AI apps through our AI for Good film.

Throughout the AI for Good course we have utilised Creative Commons licensed images from Better Images of AI’s library. Their range of images aim to avoid the misleading anthropomorphism of AI – where AI is incorrectly illustrated as a living entity with human characteristics rather than a technology created by humans. In the course eWorkbook, we have been very careful to avoid anthropomorphism in text-based descriptions of AI, ML and GenAI as well.

SEAME stands for the focus areas that AI education resources generally focus on: Social and ethical (SE), Applications (A), Models (M) and Engines (E). With other courses offering exploration of models and engines to older age groups, we focused on inspiring 11 to 14 year olds to apply the power of AI to app design whilst improving their AI literacy.

The AI for Good course focuses on students (aged 11 to 14 years) learning about ML within a real-world context. Students will explore the social impact of AI and then go on to develop ideas of a prototype AI app for a community that they care about whilst addressing one or more of the following UN Sustainable Development Goals:

• Goal 1: No poverty
• Goal 2: Zero hunger
• Goal 3: Good health and well-being
• Goal 4: Quality education
• Goal 5: Gender equality
• Goal 10: Reduce inequalities

It is essential to raise students’ awareness of the ethical issues surrounding AI through information, examples and discussion. And whilst ML models are not currently on the computing curriculum and not all students will go on to become programmers, Apps for Good believes it is critical that all young people are engaged, informed and responsible citizens of technology. We want to ensure all young people – especially those experiencing disadvantage – have the tech innovation skills to thrive in an ever-changing world. We believe that improving students’ digital and AI literacy will help to narrow the digital divide.

We prepared concise explanations of AI-related words and terms using age-appropriate language and provided examples of mobile phone features used in AI apps – all with the support of partners at Google, DeepMind and SAGE as well as consulting with teachers. Interesting though they are, we inevitably had to leave some terms out such as the AI effect (when an ML model is successfully integrated into day-to-day processes we tend to not think of it as AI). We wanted to maintain a laser focus on the key areas of AI for KS3. In further addressing Waite’s concerns, the AI for Good course includes clear AI-focused learning objectives and associated assessment activities.

GenAI models draw on a wide array of text and image-based data which makes it impossible to guarantee age-appropriate outputs. In view of this, companies and organisations have erred on the side of caution by introducing an age limit on GenAI which is 13+ tools (some also require parent/guardian permission). We believe it is critical for young people to evaluate GenAI tools in the classroom but instead, for now, we have had to simulate the use of GenAI. We’re now waiting for the development of a suitable, student-friendly GenAI tool which we can incorporate into the resources.

Across three of the four nations of the UK, we have received positive responses from educators, some of whom have undertaken the pilot AI for Good course.

Teacher Jaime Beaumont, Head of Computer Science at Stoke College, England, who has delivered Apps for Good’s pilot AI for Good course believes the course is “really good… I love how you explained AI”. Beaumont believes “This course will help students develop a deeper understanding of AI and the role it will play in their future. It will get students thinking about potential benefits for society and I’m sure it will have a positive impact on many, as it encourages them to perceive themselves as catalysts for change, empowering them to become developers and contributors in this evolving landscape.”

Tomi Rowlands, Supporting Improvement Advisor – Digital for GwE (North Wales regional school improvement service) states “Wales is currently undergoing curriculum reform and the introduction of the curriculum in Wales which gives equality to all areas of learning. What makes the resources stand out in supporting our schools is their ability to bring current and relevant issues in the world and get learners to understand them and do relevant tasks in code.org to further their digital competency.” Rowlands continues: “the moral issues with AI is also a massive strength of the resources which help towards Strand 1 of the DCF which is a hard one to get schools to address.”

Fraser Christie, Principal Teacher of Computing at St Paul’s RC Academy, Dundee, Scotland has also delivered the pilot course. He describes the course content as "very engaging and interesting for the pupils”, and says that while young people are aware of AI, the curriculums are a few years behind. He states that Apps for Good “have done the hard work for us with the AI for Good course, and made me think about bringing something more relevant into what we’re doing”.

Christie is “looking for materials that will allow me to confidently teach informative lessons on AI and the ever-increasing part it will play in students' lives. (The) materials will help my pupils understand how the technology works and provide them with the tools such as critical thinking to understand how to use AI responsibly, creatively and productively.” Following the pilot, he commented that the course “helps students question what they're looking at online” and said, “It's given them the skill of discernment. If I hadn't been trialling the AI course I might not have touched on this with them.”

We’re looking forward to schools across the UK delivering the AI for Good course. Book a 1:1 session with Kathy, our Education Community Manager, who can help you get started or email kathy.sheppard-barnes@appsforgood.org. The course will help you guide your students to develop a prototype AI app that focuses on a community they care about.

With thanks to Dr Emma Posey, former Senior Learning Manager, Apps for Good (@emmaposey)

We are very grateful to the following experts for their time and guidance in helping us refine our communication of the AI for Good course: Tom Cannon from Google, Dan Zheng from Google DeepMind, Dr Tracy Gardner from Flip Computing as well as Nabil Abu-Hadba, Boitumelo Matlala, Enric Soler Rastrollo and Manavpreet Thiara from Sage.

The following individuals and organisations’ work was also instructive: Alister Black from Adobe, Simon Peyton Jones from Computing at Schools, Dale Lane from Machine Learning for Kids and Azim Shariff from Moral Machine.

Apps for Good is constantly striving to improve its courses. We always welcome feedback. Whether you have tips of your own to share, have spotted a mistake, want to let us know that you enjoyed the resources or would like to share any other thoughts, please email us at education@appsforgood.org.