Artificial Intelligence (AI) has actually been around for decades. But only recently has it become an accessible tool for day-to-day tasks on our smartphones, picking strawberries, predicting crimes, and even English Language Teaching!
What is Artificial Intelligence?
A good working definition, which applies to the general understanding of AI, is Artificial Intelligence involves machines or computers that work and react like humans do.
The AI that is prevalent in current technology is what we’d call Narrow AI. That is to say it has a narrow focus and is very good at one or two jobs. For example, Shazam is very good at letting you know what song is playing, by listening to a snippet of the music, but it can’t tell you how many tickets are left for the theatre tomorrow night, or how tall Charles de Gaulle was.
An easy mistake to make is to imagine AI as machines that can think like humans do. This is called General AI, where a program could theoretically turn its hand to a variety of problems and, after a bit of observation, take up the task itself, as a person might be able to.
This kind of multifaceted intelligence isn’t replicable in computer programs (yet.) As Dr. Hadar Shemtov, Director of Research at Google, said in response to a question from the audience at an OUP dictionaries summit, “Computers are actually a lot dumber than the average guy thinks… they still need to be told what to do in the vast majority of cases.”
How does it work?
There are two basic ways AI can work. The first is rules based – the program is given a set of rules and it keeps applying these to the problem to find an answer. The second is learning based – the program observes, finds patterns and matches patterns independently.
The difference between these two is that one is taught, whilst one is learned.
This is why AI is mentioned with such frequency in recent years. Previously, AI was all rules based – meaning its capabilities were restricted by the rules that could be defined by programmers. Modern AI has shifted to learning based, making it exponentially more powerful and opening up many more exciting implications. This has become possible because the vast data sets, and the computing power to analyse them (required to help a program learn) are now available with modern technology.
So what is Machine Learning?
Machine Learning is when AI programs learn how to complete a task. Machine Learning involves a computer analysing large quantities of data, recognising patterns in the data, and drawing out conclusions or solutions from these patterns. Often, in order for patterns to become apparent, huge quantities of data need to be looked at. The more data available, the more likely the AI is going to produce an accurate answer.
The initial algorithm for how to use the data is written by a human programmer, but the computer then applies this to vast data sets. This means that AI can notice patterns and provide answers far faster than any human – and sometimes provide answers humans would have been incapable of ever producing. But, it also means that the quality of the data needs to be good. If you input a poor data set into AI, you’re only going to get poor output. If you want to see Machine Learning in action, Teachable Machine is a great free tool developed by Google.
What can we use AI for?
Well, just about anything. Uses of AI have already done great things to make our modern and connected lives more efficient and personalised. But there are also numerous potential benefits for education, which could improve the working lives of teachers and the learning journeys of students.
To date, Oxford University Press has already built products that make use of AI to personalise learning materials, we’ve partnered with Edwin to create an English tutor chatbot, and with Mobilinga to create interactive adventure stories, delivered via Alexa. We’re excited to continue exploring how we can make use of AI in education with our partners, using it to create new learning materials and content to help the world learn English.
Harry Cunningham is a Partnerships and Innovation Manager at Oxford University Press in the ELT division. He’s focused on enhancing and bringing OUP’s English Language Teaching content to life with the latest and best technological solutions.