I’ve been thinking a lot lately about Artificial Intelligence (AI). I’m excited about the technology’s potential, but I’m concerned about the direction the business could go, and I want you to think about it, too.
I’m excited about AI for a lot of different reasons. Films, such as 2001, Terminator, The Matrix, Her and Transcendence, are thought-provoking and show the range of possibilities. The augmentation of human capabilities, as described by Ramez Naam, in his book, Nexus, is another interesting view on the future. But I think it’s the potential scale of the business – on the order of the PC or Internet – that makes this topic so much fun to think about.
My concern was crystalized in an article I read recently in Wired, The Three Breakthroughs That Have Finally Unleashed AI on the World. As the title suggests, AI is coming soon, but it will be presented to you and me like a utility served up by Amazon, Google or IBM:
Amid all this activity, a picture of our AI future is coming into view, and it is not the HAL 9000—a discrete machine animated by a charismatic (yet potentially homicidal) humanlike consciousness—or a Singularitan rapture of superintelligence. The AI on the horizon looks more like Amazon Web Services—cheap, reliable, industrial-grade digital smartness running behind everything, and almost invisible except when it blinks off. This common utility will serve you as much IQ as you want but no more than you need. Like all utilities, AI will be supremely boring, even as it transforms the Internet, the global economy, and civilization.
I find the view that AI’s power will be centrally-controlled like a utility, very troubling. Kevin Kelly, the Wired author, is implying that one company or a small number of very large companies can be trusted to create and maintain a singular monolithic standard for AI.
I recently came across the new Domino’s Pizza ordering app, “Dom,” and as crazy as it sounds, I much prefer Dom’s view of the world to Wired’s. Watch the Domino’s 30-second spot. In the commercial, the guy is asking his digital assistant for advice on all sorts of topics. The assistant only knows about one topic. When he finally asks for a pizza, Dom quips, “I thought you’d never ask.”
Why shouldn’t we get our AI from our preferred expert in that field? Why should we rely on Google or Amazon to be expert in all fields? Even in pizza, we may prefer thick crust one day, or thin crust another. Aren’t the values of expertise and intelligence really found mainly in the eyes of the beholder, and don’t these values morph and improve over time as we learn and share more information and experiences with each other?
From the business model perspective, wouldn’t you prefer a world of many competing ideas and services, rather than a few mega-providers? While I sometimes like the convenience of Costco or Amazon, I would find life supremely boring if that was the only place I could shop.
Or think about healthcare. If I’m in pain, I might get very different diagnoses from a Western doctor, a Chinese doctor or an Ayurvedic doctor. Will the Wired model give me viewpoints from all of those medical disciplines, or will its algorithms figure out what’s best for me and filter out the other choices?
I keep coming back to the idea of The Matrix, an all-knowing or all-powerful AI that ultimately lays waste to humanity, because it has no diversity of ideas and is separated from human thought.
In my view, AI-driven intelligent agents should be experts in narrow fields, can share or sell their expertise in real-time micro transactions, and will be subject to the control of human arbiters of right and wrong. My ideal model provides increased sharing of diverse sources of knowledge, with the self-regulatory model of biological diversity, including, and centered-around, human minds.
What can you do to help? Shop small and local. Use Dom instead of Siri. Trust your eyes, ears and gut over the smartphone. Meditate. Help ensure intelligence in the future is artificially augmented, but doesn’t become artificial.
In Part 2, I’ll explore ideas for creating a viable alternative to the monolithic model of AI. Happy Holidays!