At the same time, our understanding of how we personally interact with AI in our day-to-day lives, and how we can use it to our advantage, remains limited. Ask a class of 13-year-olds what they think of when they hear ‘artificial intelligence’, for example, and the answers tend towards a common theme: ‘creepy’; ‘sinister’; ‘taking over the world’. How do we retain interest levels while grounding AI in reality and preparing students for the workplace of the future?
For young people to be able to lead and succeed in the data-driven economy, a strong understanding of this ever-evolving technology is paramount. In order to engage students with this topic, lessons should not only highlight the many forms that artificial intelligence can take in the real world, but also offer tangible experience of and interactions with the technology. Here are just some of the angles from which this topic can be approached, and suggestions for resources that can complement them.
AI in action
With the recent proliferation of smart speakers and virtual assistants, this technology can be a useful framework for an initial discussion around the key tenets of artificial intelligence. Most young people will have been exposed to these devices in some form; fewer, however, are likely to identify them as an example of AI. You could ask students:
What does AI look like?
What does it sound like?
To what extent should it mirror human behaviour?
Google’s Duplex AI assistant is a great example of the capabilities and potential of this technology. In a popular video of the assistant in action, it is heard making calls to a number of different businesses, sounding sufficiently humanlike in its interactions to fool the real humans on the other end of the line. The somewhat unnerving potential of the AI is likely to hook students, while it remains grounded in reality as an aid and time saver, rather than a replacement, for humans.
Quick, Draw! is a great resource for highlighting to students another manifestation of AI. The game challenges the user to create a series of doodles, while a neural network attempts to guess what they are drawing. Coupled with its hands-on, accessible nature, this activity has the potential to engage even the most reluctant of students.
Not so intelligent?
At the same time, we need students to be critical in their appraisal of AI technology. The mantra that ‘Machine learning is written in Python; artificial intelligence is written in PowerPoint’ is a great starting point for this discussion. You could ask students what they understand by this. The aim here is to draw out the idea that machine learning can already be seen in action in industries across the globe, while AI arguably remains a theoretical concept. Has anyone truly created an intelligent machine?
There is a wealth of resources that we can draw on to assist students in forming their own opinions in this debate. The Turing test, for example, is an important concept for students to understand and remains a useful benchmark against which to measure the capabilities of AI technology. Encourage your students to read aloud some of the transcripts for entries to the most recent Loebner Prize, an annual Turing test competition. Would any of these convince them that they were speaking to a real human?
On the website AI Weirdness, meanwhile, research scientist Janelle Shane publishes the entertaining results of her experiments training neural networks on existing content across a range of topics, from cat names to knitting patterns. Taking Halloween costumes as an example, you could allow students to explore; with suggestions such as ‘sentient stone’ and ‘a skunk in a moose suit’, it should quickly become apparent to them that this technology has some way to go in capturing the uniquely human traits of creativity and humour.
AI and machine learning can also be ideal starting points for generating lively debate around other key topics in Computing. An unplugged activity in which students create their own algorithm to guide a visitor from the school reception to their classroom is an ideal catalyst for a discussion about the differences between how people and machines make sense of instructions. The ability of humans to apply common sense when determining a course of action can be highlighted as a strength that machines are unable to emulate, which can encourage students to understand AI as a tool to complement us rather than compete with us.
Meanwhile, developments in self-driving vehicle technology presents a unique opportunity for students to explore ethics in the context of computer science. Moral Machine, developed by the Massachusetts Institute of Technology, is an interactive tool that asks the user to judge the most acceptable outcomes of a series of moral dilemmas faced by a driverless car. By engaging with this modern take on the classic trolley problem, students develop a deeper and more personal understanding of the ethical challenges surrounding artificial intelligence.
These activities have been popular in our partner schools, with teachers commenting that pupils were “fully engaged” and found the topic “really interesting”. By providing students with an interactive forum in which to discuss and explore AI, we have an excellent opportunity to support the next generation in confidently claiming their place in the modern world.
Get creative with data
AI and machine learning form a key part of Get Creative With Data, a KS3 (ages 11–14) data science course from the David and Jane Richards Family Foundation (@DJRichardsFF).
Complete course materials are available for free to all state schools. For more information about introducing the course at your school, contact [email protected].