The work we do at Apps for Good
It’s plain to see that the prominence of machine learning is only growing exponentially, and the global economy needs people that understand it well. ”
Our product manager Fergus talks about the need for young people to become the next generation of social leaders in the machine learning field.
A few years ago I was on the underground coming home from the Apps for Good office and I decided to put a bit of music on my phone. I’d been using Spotify for a while and earlier that day I had looked through my “Spotify Wrapped” for the year. One of the things it had pointed me to was the “Tastebreakers” playlist it had created for me, a list of one hundred songs that I hadn’t listened to on Spotify before but that it thought I would like.
I remember listening to the first ten or fifteen songs on that journey and being blown away by how personalised it felt. It literally did feel as if someone had sat me down in person and talked me through the kind of thing I liked, coming up with songs that I knew but hadn’t listened to in years or some that were completely new.
It took a few seconds to realise that nobody at Spotify had made this playlist for me per se - a Machine Learning algorithm had not only taken into account the music that I listen to, but collected data across millions of listens from other people and organised it to make a decision about which 100 songs I would like.
At the time I was managing Apps for Good’s partnership with Spotify, and I took this note down on my phone about how we could use their in house expertise to support our Machine Learning students.
Industry partners like Spotify and SAP have worked with us over the years since to bring a strong understanding of the capabilities of Machine Learning to young people.
They have helped us to build courses and resources, boil down and explain complicated technical concepts in a way that young people can easily understand, and help put the technology into a context that truly matters to students engaged with Apps for Good. But why is it so important, now more than ever, for them to understand this?
In order to answer that question I think we should quickly look at what machine learning is, and how it is used every day.
What is Machine Learning?
As adults, we are able to process enormous amounts of information in a single second. What machine learning provides is an augmentation to this.”
Machine learning is a subcategory of artificial intelligence focused on robotic algorithms making increasingly intelligent and complex decisions based on increasingly large amounts of data.
This sounds complicated from the outset, but it is a very similar model to the basic development of the human brain.
When a baby is born they are essentially an empty machine learning model with unfathomable potential. As they pick up information through sensory experiences they develop the ability to make complex decisions and analyze things in more detail.
As adults, we are able to process enormous amounts of information in a single second. Where machine learning comes in is to provide an augmentation to this. Computers are capable of analysing sets of data in a fraction of the time humans can, but just like humans they learn by spotting patterns and themes.
This is how technologies such as facial recognition or spam filters on our emails come about. The smartphones we all now use are constantly getting better at identifying our faces from different angles and in different lighting in order to provide a seamless experience without compromising on security.
So why is it so important for young people to understand the possibilities of this technology now? Not just because of its potential to build groundbreaking technologies that can change the world for the better or the need for experts in the field in the job market of today, but also it’s potential to cause problems.
The Ethics of Algorithms & Machine Learning
So much social and political power can be commanded by algorithms, a good reason why we build social responsibility into the core of our programmes.”
Anybody in the UK who works in the education sector, in a school or is a parent of teenagers at schools will have seen that an algorithm has made big news. Due to the Covid-19 pandemic, traditional exams had to be shelved, and the government's solution to this problem was to look at a broad spectrum of academic and crucially social data to figure out which students received which grades.
This meant that the social demographic, ethnic group and gender of any given student was taken into account within the algorithm, reducing the grades of exceptional students across the board to fit with a broader average, but particularly affecting those in more challenging circumstances.
This is a perfect example of how much social and political power can be commanded by algorithms, and it’s a good reason why at Apps for Good, we build social responsibility into the core of our programmes.
However, one of the alarming things about this algorithm is that it’s actually “relatively dumb” and not powered by machine learning or artificial intelligence. It was simply a complex statistical model.
With our machine learning course, we highlight how smarter machine learning algorithms are quite literally being developed to make life or death decisions. Driverless cars are just around the corner, and our course contains an exercise from the "Moral Machine" that highlights some of the very human decision making that we are now asking machines to make for the sake of automation.
Whether it’s deciding between an A* and a C at A-level for students from a disadvantaged background with a model powered by stats, or a decision to avert danger for a driver or a pedestrian using an extraordinarily complex machine learning model, the ethics of the subject are a crucial part of what young people need to learn in order to become responsible leaders in AI.
How do we do it?
Students design their own machine learning model that solves a problem that matters either to them or to their community. ”
The resources we have created take young people through the basics of what the technology is and does, the importance of the ethics surrounding it, and then through a process of designing their own machine learning model that solves a problem that matters either to them or to their community.
We have a large community of industry experts in the subject that can help students and teachers along the way through various systems of support and advice, and have built a library of useful videos and external resources to help students on their machine learning journey.
One of the core external resources is Machine Learning for Kids, a tool used to create basic recognition systems using simple algorithms where students can easily create their own model. It was created by Dale Lane at IBM, a friend of Apps for Good and Machine Learning expert, and is an invaluable tool in bringing the course to life.
Power in the hands of young people
Putting machine learning in the hands of young people gives young people a tremendous amount of power to wield when it comes to shaping their future.”
It’s plain to see that the prominence of machine learning is only growing exponentially, and the global economy needs people that understand it well.
However, this is not the primary reason why Apps for Good advocates for young people learning about it. Apps for Good’s mantra has always been about young people using technology to change their world. Putting machine learning in the hands of young people gives young people a tremendous amount of power to wield when it comes to shaping their future.
I get up every day to work towards improving our product at Apps for Good in the hope that one day, one of our former students will create something that will give me the same eureka moment that I had on the tube back in 2018.
Something as simple as reminding someone of some songs they used to love but had forgotten the name of can make people happy on a global scale, and when applied to even bigger issues such as mental health, climate change or perhaps even a global pandemic, the sky really is the limit in terms of what the future holds if we can impart those skills to young people right now.