Crapbots: How Fake #AI and not-so-deep Learning Could Stunt the Bot Revolution


Posted By John Koetsier (Source)

Yesterday I created a bot.

Not a very smart bot, and not a very useful bot. It doesn’t feature any artificial intelligence, and it won’t learn from any interations with visitors, chatters, or customers. But I am now officially a bot developer, sort of, thanks to Motion AI.

bot ai Johnbot

Meet “Johnbot”

Earlier this month Facebook CEO Mark Zuckerberg said that more than 10,000 developers are building bots for Facebook’s Messenger platform, a platform that he clearly wishes to become the backbone of a new way for millions of companies to interact with billions of current and potential customers.

This new means of interaction is one that Chris Messina, inventor of the hashtag and current head of developer relations for Uber, calls continuous partial productivity. You chat with early-messenging-tech adopter KLM, for instance, about a flight to NYC. You continue working, or chatting with a colleague, or watching the game while KLM’s bot goes off and checks dates, times, prices, and connections. It comes back to you with a few options, and in between opening a beer or working on a spreadsheet, you pick one. The bot does its thing, and you get your tickets … plus maybe a hotel reservation, a seat at the Mets game, and dinner reservations.

That’s the future that bots, personal assistants, and artificial intelligence hope to deliver to us. It’s something Microsoft is working on with Cortana, Apple with Siri 2.0, Google with Google Now, Amazon with Echo/Alexa, and Facebook with Messenger and maybe a little dash of Zuckerberg’s pet project du jour: building Jarvis, the AI from Iron Man.