Rohit Prasad, the scientist in charge of Alexa’s development, recently gave MIT Technology Review’s Karen Hao one of the most terrifying interviews in modern journalism. We know how dangerous it is to let bad actors run amok with AI and our data – if you need a refresher, recall the Cambridge Analytica scandal. That’s not to say Prasad is a bad actor or anything but a talented scientist. But he and the company he works for probably have access to more of our data than ten Facebooks and Twitters combined. And, to paraphrase Kanye West, no one person or company should have all that power. Hao writes: The idea of Alexa being an omnipresent companion looking to orchestrate your life should probably alarm you. But, for now, the work Prasad and the Alexa team are doing isn’t scary on its own merit. If you’re one of the eight or nine people on the planet who has never interacted with Alexa, you’re both missing out and not really missing out. Virtual assistants, today, are equal parts miraculously intuitive and frustratingly limited. With one interaction, you’ll say “Alexa, play some music” and the assistant will ‘randomly’ select a playlist that touches the depths of your soul, as if it knew better than you did what you needed to hear. But the next time you use it, you might find yourself in a three-minute-long argument over whether you wanted to listen to music by Cher or purchase a beige chair (with free two-day shipping). Hao’s Tech Review article continues: This transition probably won’t feel like a giant leap in technology. It doesn’t take an expert to assume that the pushy version of Alexa will still stumble, struggle, and fail to accomplish language tasks that a four-year-old would understand. Everything changes the moment our relationship with a collection of algorithms goes from “Hey Alexa” to “Yes Alexa?” Read next: A beginner’s guide to the AI apocalypse: Misaligned objectives
