At 7 a.m. this morning, a robot told me the weather.
It wasn’t audible, although my phone chimed. It was The Weather Channel bot on the messaging app Kik. A message popped up, and then I asked for the forecast. I asked about the weather in Cupertino, and the robot balked. (You can only ask about local weather conditions.) So much for The Terminator taking over.
As we all learned recently when the Tay chatbot on Twitter became racist and offensive, chatbots are not that intelligent yet. However, there’s hope for the future. Eventually, an army of chatbots will take over Facebook, answering customer support questions as though you are speaking to a human, directing you to online products, and even chatting about politics (through CNN, natch) and the weather.
At the F8 conference yesterday, Facebook announced a new platform for developers to build chatbots that interact with people from the Messenger app. There was a hint of how this will work with the M chatbot that was available to people in San Francisco. You could start an initial conversation with an A.I. chatbot to order flowers before a human took over and assisted.
In the near future, fully robotic chatbots will allow you to order flowers through the entire process with no human involvement. They’ll be able to answer questions, find the best product in an online store, and handle customer complaints. Of course, speaking to a digital assistant is not new -- Apple Siri, Google Now and Microsoft Cortana have been pretending to be humans for some time.
What’s new this time is that the chatbots will be much more specific. An army will take over that helps with specific products and services.
There's a simple reason for this. Facebook needs to dominate on phones as much as the company has dominated on your computer, and chatbots can interact with you quickly through the Messenger app without much effort on your part. As an example, the Kik chatbots tend to present responses so you don’t have to type anything. Facebook bots will likely support text-to-speech soon, and offer similar features where you can tap less and get help faster.
Here’s how this could quickly evolve. When chatbots get more specific to brands, they can be more helpful to us in our daily routine. Say your new Audi A4 has a Facebook chatbot. The car already lets you tell people you are driving to a meeting by selecting an option in the touchscreen. A bot could tell you when a tire is low. You could ask detailed questions, like whether a certain stroller brand would fit in the trunk. The chatbot would understand context. You could tell ask an Audi chatbot to find a gas station and, unlike Siri, it might direct you to one that has the best price on high-octane fuel (which is what Audi likely recommends).
Or, say you have a return to make at Target. A Facebook chatbot would be right there in the social media platform, not at Target.com. You wouldn’t need to register or follow some protocol. Maybe you’d just give your name and the chatbot would look you up. The chatbot might tell you whether the store has your size or recommend other products. It could remind you about store hours. By the time you arrive, the Target employee would already have the size you need at the desk.
The main reason chatbots will take over is that we like the idea of speaking to a robot to handle mundane chores like finding a gas station or returning a product. We’re used to this arrangement -- texting questions and getting answers, typing in a browwer and getting instant results. There’s no fuss. Chatbots are never cranky. If you have used the Kik chatbots, you know the replies are instant. Also, you may want to act like an Internet troll and start yelling, but a chatbot is always 100% patient.
The chatbots are coming. We’ll see how long it takes for them to start helping us.