Nisha Talagala (Pyxeda AI), Tara Chklovski (Technovation)
Years ago my (then) 10 year old niece asked me what I do. I tried to explain to her that I work on this super-exciting thing called Distributed Software. She asked me – “When I go to Google and search for something – does what you do help me? After I started working in Artificial Intelligence (AI), I am much better able to explain to my now 11 year old daughter what I do. My daughter and her friends can already see how AI impacts their world, and thanks to advancements in AI technologies, they are already able to build their own AI to solve problems they see around them.
There is growing awareness that AI skills are critical to the next generation of global economic development. AI can also help increase a girl’s interest in STEM. This article explores the unique attributes of AI that, if properly harnessed, could serve a dual purpose: help prevent the STEM gender gap from continuing into AI, and use AI as a positive force for STEM interest.
What is AI, and how is it related to Programming, Robotics and STEM?
AI enables computers to do some tasks that human brains do: recognize images, understand text, find patterns in examples, create game strategies, etc. AI programs usually start by Training (or Learning), and then apply the learning to solve new problems, or Predict what is going to happen next.
For example, say that a human asks a robot to go to another room and pick up a cup, saying — “Robot, please get me a cup from the kitchen”. The AI is the software that talks to the human, understands his/her voice, understands what he/she asking for, knows where the kitchen is, knows where it currently is, and maps a route from its current location to the kitchen and the cup. Additional Robotics and programming can move the robot to the kitchen and to the cup location (following the map computed by the AI), lift the arm and get the cup.
AI can work alone (as with auto-correct on your phone) or with Robotics as in the example above. It can also involve programming (for example in smart applications) and has a lot of underlying math. However, one of the best things about AI is that, with today’s technology, you can use AI to solve real problems without knowing much (or anything) about the math or even doing any programming up front.
Why AI for Girls? Why Girls Particularly?
AI is of course suitable for all kids, but has elements that are particularly appealing to girls. Many real world problems can and do use AI. This brings the technology closer to the girl. AI has lots of tinkering, but the tangible benefits are clear. For example, in its first season running an AI program for families, Technovation, a global tech education nonprofit, found families were eager to learn and apply new technology, like AI, to problems they experience in their communities. For example, a mother-daughter team from Palestine created an AI-based tool to help parents and caregivers recognize signs of bullying and violence. They created an AI image recognition technology that analyzes kids’ drawings looking for abuse indicators, and notifies a trusted adult/professional.
AI can also be reachable to girls in ways not historically possible with STEM. For example, with tools like AIClub, it is now possible for school age kids to build and use an AI with only a web browser and Wifi , making AI reachable within the home. There is no need to go out and buy extra hardware. Broader initiatives to connect households can be leveraged for AI reach. Almost 60% of the world’s population is expected to have been internet connected by the end of 2019. Finally, AI can be almost entirely on the cloud – meaning that low cost tablets, computers, or even smartphones can be used as access points and anyone anywhere can do even heavyweight AI.
Many AI-For-Good initiatives show girls how AI can be used to help people and solve problems. For example, through 14 years of running Technovation programs for girls and families, we have found that participants’ have a greater likelihood to persevere through challenging curriculum about complex technologies, like AI, when the lessons are taught through the lens of a problem they personally care about. In an age of gamification, this is especially true for girls. Giving girls tools to collaborate and work together to solve problems is an opportunity to build their confidence and also address major issues of our time and theirs.
AI naturally leads to programming and math. In a survey of parents who participated in an AI competition, more than 90% believe their child developed a sustained interest in AI, 90% believe they know the prerequisites for their child to pursue an AI/STEM career and nearly the same percent believe their child is now capable of creating an AI model in the future. In another study, 100% of students who took an AI class reported a greater interest in programming, and 42% reported an increase in math interest.
The Flip Side – How girls can help AI
As AI expands, conscious and unconscious human biases are making their way into AI products, with high profile companies struggling with AI Bias issues. Since bias can enter an AI in many places (from the data, to how the problem is defined, how the algorithm is evaluated to name a few), one key way to reduce AI bias is to diversify the humans involved, bringing new perspectives into all stages of AI development. The future success of AI will require that more girls and women be part of its creation. Women are already at the forefront of this challenge, raising awareness of algorithmic bias, developing ways to remove bias during AI development, and setting industry standards for AI accountability.
A (Limited) Window
Experts describe the AI wave we are experiencing as a “Cambrian Explosion”, invoking the same kind of disruption that occurred in evolution history. Others have described AI as “the new code” or “the new electricity”. As every economy on the planet embraces AI, it is in everyone’s interest to ensure that the STEM gender gap does not continue, or even worse, expand, into AI. The STEM gender gap is a complex problem with many historical and cultural factors. We have a limited opportunity to stop these issues from seeping into AI.
It is possible to not repeat the past:
- While there were strong female role models in the early days of computing – somewhere along the way the gender gap persisted and grew. AI is a new field with already many examples of strong female AI role models. Now is the time to ensure that we establish and grow these early gains.
- New tools like AIClub are bringing AI to the reach of girls everywhere. With just a web browser and Wifi, a 10 year old can build and use an AI in 15 minutes. Progressive learning of more AI, math and programming can follow. Using tools like AIClub, educational institutions can bring AI to their programs with a top down approach where AI can be applied to real world problems by anyone anywhere.
- Community drivers – like Technovation’s families program, which focuses on real-world problem solving using AI, encourage girls and families to learn and apply AI. In a future workforce where humans are working alongside machines, we need to provide opportunities for adults and children to explore, be creative, and embrace lifelong learning – whether that is acquiring new skills, or reevaluating existing ones.
While there has been tremendous activity in AI over the last decade, as a technology it is still nascent when compared to its future potential. AI is shaping up to be the next code, and as software development revolutionized all industries, the next wave is the broad use of AI to harness insights in data. While the challenges of closing the STEM gender gap should not be taken lightly, with the tools we have today, it is possible to limit the gender gap for AI from this, our “Day One”.
Let us not allow the opportunity to pass us by.