Kids using machine learning interface
A child-friendly interface for IBM Watson helps guide kids to create machine learning models in Dale’s IBM Activity Kit.
Kids using machine learning interface
A child-friendly interface for IBM Watson helps guide kids to create machine learning models in Dale’s IBM Activity Kit.

What started as a coding tool that Dale Lane created for local schools near his home in the UK has turned into one of the most used IBM Activity Kits ever—by some estimates, as many as several hundred children use the kit every day.

Machine Learning for Kids, also called Train your Computer with Machine Learning, has been translated into 12 languages—significantly amplifying Dale’s impact beyond his local schools.

Dale, an IBM developer based in Hursley, UK, was awarded a 2018 IBM Volunteer Excellence Award—the highest form of volunteer recognition given by the company—which also earned a USD 10,000 grant from IBM for Solent Youth Action, an organization he supports.

IBM Volunteers spoke to Dale about his experience as a volunteer and the creation of his immensely popular activity kit.


Was Code Club your first volunteer experience with STEM for children?
No, I've done things like this for many years, since being at University and also throughout my time at IBM. I've regularly helped run IBM STEM events that included preparing and leading STEM activities. Outside of IBM, I've also worked with other organizations, such as serving as a governor for my local primary school, and as a founder and trustee for a local youth volunteering charity.

What do the students do in the machine learning activity kit you created?
They train machine learning models, using a simplified child-friendly interface for IBM Watson that guides them through the process of a machine learning project. They end up with a model on a few existing educational coding platforms (through integration with Scratch, App Inventor, and Python) for creating their own artificial-intelligence powered apps and games.

It's accompanied with 27 project worksheets: step-by-step recipes that give students instructions for how to train and create a project, each based on a different real-world use of AI.

It’s available in English, Chinese, German, French, Japanese, Korean, Dutch, Brazilian Portuguese, Sinhalese, Spanish, Swedish, and Turkish.

Why do you believe kids should use and understand machine learning?
Machine learning systems are all around us. We all use, interact with, or are affected by, machine learning systems every day. If our goal is to help children to understand how the world around them works, then understanding machine learning is an important topic.

Also, the impact machine learning systems will have on our lives will only continue to increase. There is a growing need for society to make decisions about how we want such systems to be managed, controlled and regulated. Understanding machine learning is an essential topic to help the next generation engage in that societal debate.

What did you do to create the kit?
Quite a lot [laughs]. There’s a detailed answer to that question, along with a timeline and much more on my blog.

The short version is that I made the tool first for local schools, not as an activity kit. It spread, mostly through word-of-mouth amongst teachers and code club leaders. After increased adoption, I contributed it as an IBM Activity Kit.

How long did it take to create the activity kit?
It's hard to say, as it's been very open-ended. I'm still continuing to work on it now. And it's an iteration on things I've been working on since 2016. It was a lot of work, many hundreds of hours but it was something I thought needed doing. And, as far as I could see, no one seemed to be doing it.

What IBM resources or people did you draw on for guidance?
Lots, especially colleagues in IBM Watson and in Design. Some of the wonderful people include David George, Simon Burns, Brian Hulse, Malcolm Cloudwell, Andrew Daniel, Adam Roberts, Bethany Simpson, Hannah Berrisford and Nivedhaa Muthu. Certainly others, and people who provide valuable input in online comments. Also, Mark Wakefield and Dominic Nolan on the IBM Corporate Citizenship team helped arrange opportunities to present early versions to teachers to get their input.

We tested the activities in many ways, including trying them out at schools where I have a long-running relationship, inviting classes to IBM offices to try them, and running the activities at local events such as Raspberry PiJams.

What skills from your work at IBM helped with the creation of the kit?
I worked in IBM Emerging Technologies from 2008 to 2011, and in IBM Watson from 2011 until 2017. In that time, I learned a lot about machine learning and how it is applied by our clients. That experience informed a lot of the topics that are covered in the project worksheets.

One of my roles during my time in IBM Watson was enabling developers to use Watson services in their applications. The objective was to take deep complex technologies and make them accessible to developers without a background or expertise in machine learning. Machine Learning for Kids is sort of an extension of that work. Some of the same principles of how to make it simple, and guide users through a process for training an effective machine learning model are consistent for both.

What is the ideal student age for the activity, group size and leader experience?
It's fairly flexible. I've run it with small groups of a few kids. I've run it with clubs of a dozen kids. I've run it with school classes of 30 or so. I've run it as a speaker-led activity for audiences of 150+ kids. And everything in between. Similarly, I've used it with both primary and secondary school age children from 6 to 17.

The project worksheets are fairly self-contained, so leaders don’t need any experience. It doesn’t hurt if someone can share personal experiences and context about artificial intelligence, but it isn’t necessary. There are plenty of people who have run the activities with no personal technical skills in AI or machine learning.

How is the kit evolving?
I've continued to update and refine the activity kit since its first release. Some of this has included adding major new features and capabilities. For example, last month I added support for new types of machine learning model, to allow support for training sound recognition projects.

Note: IBM has over sixty activity kits available to all volunteers on topics such as STEM, cultural diversity, disaster response and workplace skills.


Dale Lane is among 15 IBM teams and individuals who are recipients of the fourteenth annual IBM Volunteer Excellence Award. The award is recognition from IBM Chairman and CEO Ginni Rometty and is the highest form of global volunteer recognition given by the company to employees. It includes an IBM grant for the associated not-for-profit partner or school.

Join the conversation. Tag your tweets with #IBMVolunteers, #IBMVolunteerExcellenceAwards and follow us @IBMVolunteers.


About these stories

Read about IBMers whose volunteer efforts are improving communities around the world.


Activity Kits

IBM’s volunteer Activity Kits include everything you need for a range of activities.