The Possibilities of Artificial Intelligence in Education

This week, I had the pleasure of being invited to speak at The Index Conference in London for educators, policy-makers and head-teachers visiting our amazing City for inspiration and knowledge about how to foster creativity in children with I. T. They were particularly interested in the possibilities of Artificial Intelligence and Machine Learning. So, with thinking cap on, and just a few short hours to prepare, I was thrilled to find out that A.I in education is not the work of science fiction, but is with us right now - in action, and starting to build impact.

For hundreds of years, humans have pondered the idea of building intelligent machines. Over this time, artificial intelligence has had highs and lows, demonstrated successes and unfulfilled potential. Today, the news is filled with the application of AI and machine learning to new problems. AI impacts future industry, our global economy, education and jobs but something more than that – it has an impact on what it is even be human.

So How did we get here? In the past few years we have seen 3 waves of technology that are co-inciding:

• Rise of Cloud – giving access to computing power and information everywhere and anywhere
• Rise of Data – generating vast quantities of structured and unstructured data – video, text, tweets, images collected from millions of sensors and devices
• Mobility – this iPhone 7 in my pocket now has more compute power than the old SuperComputers of the 1970s
• These 3 waves of technology are underpinned by the need for Security & Privacy

With so much information and power at our fingertips, we needed to do something to help us control and use this to help us be more productive.

For example, there are over 8,000 papers and medical journals published a day that Doctors could read. And new medical data is doubling every 60 days. They need a system that can read and learn from all this information, and these insights need to be inserted within the doctor’s workflow

A.I. is a solution to these challenges

Some public figures, including Stephen Hawking, predict a future where humanity is wiped out by intelligent robots. Others, meanwhile, anticipate an era in which mankind is freed from the need to complete boring tasks, aided by intelligent robot assistants. 

But A.I. a not replacement of people.

It’s a partnership of man & machine. With every profession that AI systems are introduced, there is a 2-way relationship – it can help you do a better job. For instance, who would think that AI would improve the human creative process in Music?

Last year, Grammy award-winning music producer Alex Da Kid paired up with IBM to see if they could create a song together. Watson’s ability to turn millions of unstructured data points into emotional insights would help create a new kind of music that for the first time ever, listened to the audience.

Watch Behind the Scenes of 'It's Not Easy'

There is no doubt that AI is beginning to encroach upon our lives, whether it is Amazon’s Alexa controlling our homes, or automated chatbots answering our queries online.

But what about education? Will AI realistically have any impact on the way in which we teach and learn?

But maybe I’m jumping the gun here, so let’s step back and talk about what is AI?

What is AI? -

The history of modern AI has all the elements of a great drama. Beginning in the 1950s with a focus on thinking machines and interesting characters like Alan Turing and John von Neumann, AI began its first rise.  Back then, AI focused on how machines could perform any intellectual task that a human could.

It wasn’t really until the 1980s when Machine Learning – a subfield of AI - became a prominent area of research, that AI started to become a reality. Machine Learning’s purpose is to give computers the ability to learn and build models such as prediction within specific domains.

More about Machine learning

You may be familiar with a teaching tool called Scratch that is used to help to explain to children the idea behind programming. It teaches that programming is about taking a complex task that you want a computer to do, and describing that as a series of specific steps. The point of choosing from a set of simple blocks and snapping them together is to give kids an accessible visual metaphor for the concept of programming.

 Machine Learning is a bit different.

With machine learning, to get the computer to perform a complex task, you collect a set of examples of that task being done. The computer learns how to do that task from the examples that it is given, then can predict the next steps with increasing levels of accuracy.

Say that I wanted to teach you how to kick a ball. In a traditional programming approach, I’d give you a series of detailed instructions. I’d tell you

-      how far to lift your foot back?

-      what angle to put your ankle at

-      how much to bend your knee?

-      How fast to move your leg?

In effect, I’d tell you exactly what I wanted you to do, how I wanted you to do it, how fast I wanted you to do it, and the order that you should do things in. To do this based on a rules system would be very slow and need constant updating.

With Machine Learning, examples of kicking balls would be given to the system to train the computer to be able to recognise what kicking a ball looks like, Then the computer would stand a better chance of recognising how to do this well

Artificial Intelligence and Machine Learning techniques are transforming the way we approach real-world problems in almost every industry. We might not realise it, but we all use systems and applications that depend on machine learning every day.

-      Spam filters are a good example. 

-      So are assistants like Siri, Google Now and Alexa.

-      Systems that translate one language to another – trained on examples of documents that have been manually translated.

-      The auto-suggest on my phone keyboard, that suggests what word I might want to write next – trained on examples of what I’ve typed before.

-      Credit card fraud detection – trained on my buying patterns to recognise a purchase that might not actually be me. 

-      And many, many more. And this is just in the consumer space.

Examples of AI – IBM Watson

At IBM, our implementation of AI is called IBM Watson.

Watson is an advanced computing solution capable of providing expert-level insights and advice to human beings working in a wide variety of fields, including medicine, and education.

Specifically, Watson is good at analysing massive amounts of data (research publications, news, social media, organizational data, medical data, lesson plans, etc.) and providing industry professionals (doctors, teachers, accountants, etc.) with relevant and clear recommendations based on a user-level request.

IBM Watson made its debut with a state-of-the-art question- and-answer interactions on Jeopardy Quiz Show.

To be clear – what IBM Watson does, is not just Search.

The Jeopardy Quiz show is an open domain of questions and answers. Key to being successful in the show is the interpretation of Context – where contestants are provided the answers to a question, leaving them to identify the question itself. To take an answer, follow that thought back up and parse it to the question, this is not search – this is context.

More than 10 years later, we work in a world of computing that is not merely a series of just rows, columns and basic algorithms of numbers. We deal with unstructured data, images, sounds, language.

 IBM Watson on the Jeopardy quiz show was to show the art of possible. Now we can use it to find solutions to the unsolvable – in healthcare, education and other industries such as finance and retail. IBM has since extended Watson services and made them available on the IBM Cloud for software developers to add to their own applications (A.P.I's) https://console.bluemix.net.

Services such as visual recognition, speech-to-text, and text-to-speech function; language understanding and translation; and conversational engines to build powerful virtual agents of their own.

How does this work in real life?

In Healthcare, we have focused on oncology, the diagnosis and treatment of Cancer. Watson was fed lots of text from articles and research- over 18 million papers – then we worked with doctors to train Watson to get better at providing correct answers. We started with Breast and Lung cancer – now 80% of cancers are covered to help Doctors during the diagnosis and treatment phases with Cancer patients.

In Education, a good example is Dino, from Cognitoys. The Dino is a small plastic dinosaur that runs on AA batteries, but the real magic is that its brain is powered IBM's Watson, with extensive natural language skills. Dino is available from Amazon for just $59 - so A.I does not need to cost a great fortune! The toy answers questions from children upto the age of 9 years old and features some child-friendly customisation. It scolds curse words, refuses to answer naughty queries and can adapt it's answers to the child's age. The educational aspects to the toys include its ability to define words and teach kids how to spell.

Although Dino is a fun just for fun, recent research has found that autistic children are more comfortable interacting with robots than humans, in part because robots are more predictable and can be controlled. Experts also say teaching social skills to children with autism requires frequent repetition. Last time I checked, robots are great at repetition. So, applying AI and robotics to educational toys could prove a way to interact for these children, their educators and parents.

Why is this topic important for us as educators and parents to understand?

Every facet of our life is impacted by technology, with all elements of kids' future lives likely to be impacted by technology to some extent. The children we are teaching today will be taking up jobs which don’t exist yet – we have to prepare them for a job marketplace we don’t know and don’t understand.

Whatever their goals or career aspirations, they are going to need to know how to make technology work for them, not just be consumers of that technology. Learning about real world uses of AI and giving them the basic tools to start applying AI and ML to simple coding projects at an early age, will help them prepare for a future that they can shape.

These facts are recognised by employers, policymakers and educators with supporting programmes rolling out around the world such as Skills for Londoners, the London Mayor’s Digital Talent Programme on which I am a member of the steering committee. This programme is investing £7million to plug a growing digital skills shortage in London’s labour market with diverse, home-grown talent It aims to support young women and young Londoners from diverse ethnic and disadvantaged backgrounds aged 16-24 years old by working with industry, skills providers & schools, women in tech organisations, youth & community groups and extra-curricular providers

This programme has identified a number of priority occupations which employers are seeking in London including:

• Software and applications design and development
• Software Developers
• Games Developers
• Hardware, network, cloud and infrastructure engineers
• Data Scientists
• Digital Marketing skills
• Skills in Cyber and Information Security (
• TV and Film Production

It is no longer seen as essential that young employees have gone through University. There has been a dramatic rise in places for apprenticeships which are now at an all-time high in the UK. Young people are now taking advantage of new, high-quality apprenticeships that take learners right up to degree level and give them the skills they need to build a career. Government figures show that there are more young people starting apprenticeships than ever, with over 84,000 starts by under 19 year olds between August 2015 to January 2016.

Higher apprenticeships are delivering employers and are successfully training the next generation of engineers, computer developers and business people. Some of the largest companies in the country, including IBM, already offer higher apprenticeships and many are in designing new degree apprenticeships.

Until now we've made the mistake of thinking coding is a distinct career path, set aside for 'techies'. However, this will change. Specialist software developers will of course be still be needed in abundance, but every individual - whatever their chosen career - will need coding.

An example of this is that currently, 99% of all businesses in the UK are classified as small to medium sized enterprises. That's 5.4 million businesses. Yet, according to the House of Commons Business Statistics Report 2017, nearly 2 million of these businesses don't have an online presence. The research goes on to say that not having a website costs the UK economy some £343 billion every year. This would mean a boost in turnover of nearly £174,000 per SME, per year - if they were to introduce a website. Add an AI-enabled chatbot to that website to provide basic customer service, and you have the opportunity to increase that.

As the digital revolution takes over and kids aspire to become entrepreneurs, educating the next generation to code, especially coding websites and adding in AI capabilities, is not only good for the individual, but also for the economy.

The UK government recognises these facts, and made significant changes to the Computing National Curriculum in September 2014, with children as young as five being introduced to elements of coding.

So how can we help teachers deliver education for our digital creators of the future?

Schools which embrace AI may find the role of teachers adapting to be closer to that of a mentor , guiding children through their education. There are some interesting educational support for teachers which have recently been announced:

Teacher Advisor with Watson

Teacher Advisor With Watson is the first ever Watson-powered tool designed specifically for teachers. Starting with advising teachers on resources and lesson plans for teaching Maths.

How does this help teachers?

Because of Watson's ability to identify and then refer back to relevant concepts and relationships within a large body of information (in this case, the information we're talking about is the library of free educational resources that we uploaded to, and had enriched by, Watson, Teacher Advisor can and does return highly informed results and recommendations.

Watson presents the best-known matches between a given search term and all the vetted resources it has analysed—even if the most highly-ranked resource doesn't contain the exact keyword you entered in your search.

We began with 1,000+ lessons, activities, documents, and videos for Watson to ingest. Then, using Watson Discovery Service all of these resources were organized, studied, and enriched by Watson.

Because it understands human language, Watson can add education-related concepts to individual resources, and can also identify meaningful relationships across resources. This ultimately helps Watson make its resource recommendations. Former teachers and math experts, as well as other technical experts, provide Watson with the necessary guidance and training along the way.

We've only touched the surface of what's possible with Watson! With your help, and thanks to a continued partnership with a dedicated team of former math teachers and coaches, we plan to train Watson to better understand the elementary math domain, ultimately enabling it to provide more targeted and personalized recommendations for individual teachers.

https://www.youtube.com/watch?time_continue=160&v=Kzp3YuC_Dr8

What we should teach children and A.I and Machine Learning?

AI and ML will not replace the need to learn how to code, but it adds a new capability to the tool box. Here's how the two concepts can be moved together during sessions...

Machine Learning for Kids

Machine Learning for Kids is a new free resource from IBM that introduces school kids to the principles and implications of machine learning by helping them train simple machine models and use them to make games and interactive projects. It's been written by one of our own Engineers, Dale Lane from the IBM Hursley Labs in the U.K 

Machine learning expertise is not required. The project worksheets detail everything the children will do, with explanations for why, and examples of the real-world applications of what they do.

There are 15 machine learning projects at machinelearningforkids.co.uk/worksheets each of which has its own downloadable project worksheet containing detailed step by step instructions with screenshots and explanations.  Experience of using the educational tool Scratch is helpful, but not essential (for answering student questions, or helping them with any problems with their Scratch scripts).

Example – Make me Happy

In this project, you will make a character that reacts to what you say.If you compliment it, it will look happy. If you insult it, it will look sad

The first step is using basic Rules to make the face turn happy or sad – the restrictions of this are soon seen as students will need to type the exact words contained in the Rules to make the face change

Next, students train their computer to recognise text as being kind or mean. Instead of trying to write rules to be able to do this, they will collect examples of mean and kind words - This is called “supervised learning” because of the way you are supervising the computer’s training. The Computer learns from patterns in the examples you’ve given it, such as the choice of words, and the way sentences are structured. These will be used to be able to recognise new messages.

https://machinelearningforkids.co.uk/#!/welcome

There are many other free resources to help budding software developers, including IBM Code Patterns written by our developers providing open source code and how to guides for quickly deployable applications

Summary

Back in the 1960’s when IBM introduced the first programmable systems, we felt our role was to teach to world how it could be used. For example, we helped to establish the role of computer science in Education – something that had not existed as a discipline.

And the same goes for AI – we have a role to help you to teach the world about A.I.

Last year our CEO Ginny Romety issued a rare Policy letter on the 3 guiding principles for bringing A.I to life to help us make sure it is implemented and taught effectively:

1. It’s important that people develop trust in AI systems. Its purpose is to aid humans, not replace them. AI should extend the capabilities of humans.
2. Transparency You need to be clear as you build AI platforms how they are trained, what data was used and the business model behind the systems. Humans need to retain control of the system. You own the insights – you are not training someone else’s data. We do not believe they will have a consciousness or be self-aware         
3. Skills - AI platforms must be built with people in the industry – doctors, teachers or other professionals who will use the system

Our roles - as educators, parents, inventors and implementors - will be critical to guiding these skills, to prepare children to make effective use of technologies, to enable us to put in the right services for the right jobs and re-skill people that might be affected by this powerful technology now at our finger-tips.