Artificial Intelligence (AI) has the potential to advance education, healthcare, politics, national security, criminal justice, and so much more. A world powered by AI is within reach, and its applications will impact everyone. But experts know we shouldn’t just ask how AI can make us better: We need to consider the reverse. Data bias, lack of transparency, inequitable distribution—these fundamental flaws in AI could lead to system-wide lapses in judgment. And that means all of us in the public and private sectors should devote as much energy to educating ourselves about AI as we are to developing it.
WHY THE BLACK BOX MUST BE DEMYSTIFIED Much of that education starts with the material we use to build it: AI is what it eats. Feed an AI system biased data, and it will likely make biased decisions. In 2016, it took less than 24 hours for an AI-powered Twitter account to turn into a hateful, insulting troll. And the most advanced AI-driven facial-recognition algorithms—constructed with data from its largely white and male developers—still struggle to even acknowledge the faces of women and people of color.
Education goes beyond picking the right materials for AI: It’s also about understanding how AI thinks. Many engineers still don’t fully understand how and why AI comes to decisions— and the potential results could be alarming, if AI is tasked with firing a missile or diagnosing a disease. If we’re