Last year at the World Economic Forum in Davos, Google CEO, Sundar Pichai, called Artificial Intelligence (AI) the “most important thing humanity has ever worked on.” AI is, in his book, something that is more profound than electricity or fire. Maybe even more important than language itself.
I shared a video of his talk on my LinkedIn feed recently and it ignited conversations on my feed and with friends/colleagues IRL. We talked about tech leaders getting their predictions wrong, about taking sweeping statements with a pinch of salt, and even debating if governments around the world would agree to demilitarize AI.
I don’t know if governments – and tech organizations (surely we need to hold them accountable too?) – are going to demilitarize AI. Neither do I know if Elon Musk’s prediction (“The danger of AI is much greater than the danger of nuclear warheads”) or Stephen Hawking’s warning that “the development of full artificial intelligence could spell the end of the human race” are true.
Having said that, I’ve been working with AI since 2012. And I’ve heard my share of myths about AI during this time. Here are 9 myths, in no particular order. May they bring you a smile, if nothing else!
Myth 1: Robots powered by AI are going to take all our jobs, and will soon rule the earth.
My take: This is either a long time out (~20+ years) or won’t happen at all. For the state of the art today, maybe we should have asked Microsoft’s Tay about this, while we had a chance. Or maybe Tay’s successor, Zo?
Myth 2: AI is making us all dumber by the day.
My take: It’s had the opposite impact on me and my colleagues. It’s making us think more creatively, and at larger scale than was ever possible before. https://xkcd.com/1875/
Myth 3: AI is only for large companies that can burn cash. It is a lot of investment, for very little gain.
My take: Nope. Exhibit A – Go Moment.
Myth 4: AI algos can make sense of any data that we humans throw at it.
My take: Not even close.
Myth 5: AI = Siri/Alexa/Google. It has nothing to do with anyone else or anything else.
My take: Again, I refer you to Exhibit A – Ivy by Go Moment
Myth 6: AI is one technology.
My take: It’s not. It’s a whole cluster, largely focused around Machine Learning for practical applications today.
Myth 7: AI is artificial.
My take: It’s anything but. AI is actually based on simulating how human neurons are thought to work! Heard of Geoffery Hinton, the Godfather of AI? Check out this Hello World video where he talks with Ashlee Vance.
Myth 8: AI automatically gets better over time.
My take: If designed perfectly and in a squeaky clean data environment, maybe. In the real world, meaningful improvement still requires human oversight (supervised learning vs. unsupervised learning) and careful design. This is why Google continues to invest resources into AI principles and ethics.
Myth 9: AI is objective.
My take: Biases can and do unintentionally creep into system designs. See this explanation of how ads now use AI to match copy, images and audience.
Anyone else want to add to this list? Which ones from my list made you smile? What myths have you heard?