Classrooms today are no strangers to coding or robotics―but few classrooms in the US currently teach artificial intelligence, despite AI being applied across almost every industry. Seeing this need developing, in 2017 Iridescent teamed up with NVIDIA to develop a curriculum that would demystify AI for youth, teaching them real-world ways AI can be used for good and introducing them directly to AI tools they can use themselves.

Iridescent CEO Tara Chklovski and Joe Bungo, Deep Learning Institute Program Manager at NVIDIA, recently had the opportunity to discuss this collaboration and share what we’ve learned in the first year of running the AI Family Challenge at SXSWEDU.

Why is teaching students about AI important?

Computing skills are recognized as essential skills for nearly every field, and as AI tools and technology touch more of our day to day lives, nuanced understanding of how artificial intelligence and machine learning work will be just as critical. Additionally, AI, Machine Learning, and Data Science are all rapidly growing careers, and curriculum at the K-12 level should shift to acknowledge this.

Educators and AI experts are working to develop standards and recommendations for integrating AI topics in K-12 curriculum – notably AAAI and CSTA with their AI4K12 initiative. NVIDIA has also been working with universities to share their AI-focused teaching kits with professors to help them integrate critical AI and Machine Learning concepts into their lessons. In partnering with us at Iridescent, NVIDIA endeavored to apply that idea to the K-12 space as well.

Demystifying Artificial Intelligence for K-12 Students

We worked closely with NVIDIA researchers to break down the complex core concepts of artificial intelligence technology and products into more understandable segments, and tie them to hands-on projects. For instance, to explain one highly visible and well known (if not well understood) AI technology, we created a game inspired by self-driving cars.

Example: Self-Driving Cars Demystified

Students and families are asked to build a game that incorporates the ideas of sensors, data, and algorithms, and in the process of building and playing the game, they develop a high-level understanding of how algorithms in smart cars work. For the self-driving game:

  • Families agree on a specific action determined by pulling a playing card from a deck – for instance, if a red card is drawn, the player must turn on an LED
  • As their fellow family members flip cards faster and faster, the player who is acting as the “car” must take action more and more quickly based on the “inputs” they’re receiving from the cards (for instance turning the LED on as red cards are drawn and off as black cards are drawn)
  • Through this translation of a computer algorithm into a physical experience, students develop an understanding of sensory data and algorithms.

For the first year of the AI Family Challenge and its curriculum, we worked to demystify these core AI concepts by creating engaging hands-on projects and then reinforcing these lessons by asking families to apply what they learned to real life situations. Starting from core concepts related to computing and artificial intelligence, the curriculum builds to more complex ideas that help learners actually apply those ideas in their own lives. From lessons about circuits, families move on to lessons about neural networks and parallel processing, and then to concepts about data, bias, and identifying problems AI can solve.

In the first year of the AI Family Challenge, 7,500 participants from 13 countries joined us in learning about and creating with AI. About 70 partner sites – most led by educators – brought families together for a multi-week program to explore AI as a community. After participating,

91% of parents believed their child developed a sustained interest and burgeoning proficiency in AI and 90% of parents believed they understood what a career in AI or STEM fields would look like for their children.

91% of parents believed children were more interested in AI

Getting better at demystifying AI – What’s next?

As we prepare for the next year of the AI Family Challenge, Iridescent, NVIDIA, and other leading tech and AI companies continue to improve the program. As we refine our curriculum we want to demystify more elements of AI and are working to convey the following key points as clearly as possible:

  • AI is different from human intelligence
  • Computers perceive the world using sensors
  • AI agents have models of the world and use them for reasoning – and these models are imperfect
  • Computers can learn from data
  • It is very difficult to develop comfortable interactions between humans and AI
  • AI applications can (and do) impact society in huge ways

We want students and families to walk away understanding of these core concepts as part of their foundational AI knowledge. If you’d like to learn more about the AI Family Challenge and our curriculum, you can explore here.