Consider this: I want to book a flight from Atlanta to New York for next weekend, and want to get the cheapest ticket available. I also I want it to be with an airline that is connected to my rewards program, preferably on a flight that has wi-fi. How would I do it? I will present two scenarios:
The traditional app way:
- I open up my phone and search for flights in the American Airlines app.
- I filter it to get a cheap flight that has wifi
- Unfortunately it’s too expensive, so I open up my Delta app and repeat the process.
- I finally find a flight with a good price, I select it.
- I log in and input my credentials
- I follow the on-screen instructions and pay with my credit card
- I get the ticket and add it to my phone
This is the fastest way you can obtain a flight today. It’s not bad, but I do have to sit down to do all the process. Now entertain for me please, the following vision of an AI-based interface:
The AI way:
- I open up my phone, hold a button and say, “Find me a cheap flight from Atlanta to New York for next weekend, make sure it has wi-fi.”
- The AI replies, “I found you a couple flights. The top choice is a non-stop Delta flight that departs at 10:00am and costs $250. Would you like me to buy it?”
- I reply back, “sure!” and the AI responds, “Great! I put your ticket in your phone”
The AI interaction in this situation is superior to the traditional app interface. It’s faster, more engaging, and makes the hassle of booking a flight less so. Now, I’m not saying that every single interaction should be replaced by AI. There are indeed some instances where its better off that you see a screen and make a decision for yourself. But the fact of the matter is that having an AI is like having an assistant: it reduces logistical hassle and presents you with curated information that is relevant and useful.
Because of the superior productivity that AI provides, I believe that apps will be supplanted by AI services as the primary interface we have with our digital devices. In the long-run, natural conversation (whether it’s in speech or in a text conversation) with AI programs, designed to serve particular needs, will be the go-to approach for us to interact with our devices.
I envision a future where there will be a “general” AI installed on your device (think of Siri or Alexa), connected to multiple AI services that specialize on a particular subject eg. Finance, biology, football, etc. These specialized modules will have large corpuses of data specific to their topic and allow users to get information or perform tasks particular to that area of knowledge.
There are some who would say, “this is great! AI will take over and we wouldn’t have to work anymore!” Which I completely disagree. Automation does not equal lack of work for people. It simply means that we can better allocate our skills to other areas. Creating all of these different AI services will mean that thousands of developers will be needed to build them, and millions of people will be required to run the companies that collect, assemble and input all of human knowledge into artificial intelligent systems.
Artists, writers, thespians, anthropologists, designers and a whole host of creative people will be required to create the scripts and personalities that run these systems. Far from making people unemployed and heralding the death of humanities, the mass-production of AI services will be a renaissance for the humanities as it explores how to re-create an artificial human mind that can service its creators.
Should we become worried if these AI systems become our robot overlords, as Hollywood movies and Elon Musk portray? You’ll find out in my next blog post, so stay tuned!