Title: The Dark Secret at the Heart of AI
Source: MIT Technology Review, via Google Alerts
“If it can’t do better than us at explaining what it’s doing,” he says, “then don’t trust it.”
My Impressions: I am not going to say that machine learning is misunderstood, but the fact that most everyone uses AI and ML interchangeably says that for me.
Understanding machine learning, called deep learning by astute marketers, takes a lot of knowledge in the more complex concepts of AI (neural networks, cognition, etc.). We are not talking about what most companies are using today (advanced analytics) and calling AI – but about totally trusting the machine to make decisions without a good understanding of what it is doing. All those Sci-Fi films you saw that we use to base our (incorrect) understanding of AI are focused on ML, not the ability to process data, find patterns, and predict outcomes (which is what we do for AI, mostly, today).
This article will help you understand why Stephen Hawkins, Elon Musk, Bill Gates and other luminaries are warning about what AI can become (they mean ML; don’t want to overemphasize the difference – but it’s critical). It will also show you how my take on this, we are in control – regardless of what we tell computers to do, is valid if we do the right things going forward.
Good read, plan 10-12 minutes for it.
(Cross-posted @ thinkJar)