We’ve seen it in movies like The Terminator and The Matrix:
Once machines achieve a certain level of intelligence, they try to take over. They try to replace humanity as the dominant force on the planet.
(Fortunately for us, humanity usually emerges victorious in these scenarios thanks to the heroic actions of a few people wielding laser guns.)
But why does the “human vs. robot” narrative come up so often in the first place? What makes us so fascinated with the idea of doing battle with intelligent machines?
Here’s how I look at it:
Ever since engineers started working on artificial intelligence in the 1950s, we’ve been testing machine intelligence using our own intelligence as a measuring stick.
It started with the Turing test, which asked if a machine could ever act intelligently enough to be indistinguishable from a human during a text-based conversation. The whole point of AI at this time was to see if a machine could successfully engage in an activity that required a human level of intelligence.
Flash forward a few decades and there are now annual Turing test competitions that pit intelligent chatbots against human judges. The judges don’t know if there’s a human or a machine on the other end of the messaging app when the test starts. If the machine can dupe the judges into thinking it’s a human, it’s said to have passed the Turing test.
Of course, humans also do intellectual battle with intelligent machines in a variety of other arenas. It’s not just about mastering the art of 1:1 conversations.
In 1997, for example, IBM’s chess-playing computer Deep Blue beat a world chess champion.
Then in 2011, IBM’s question-answering computer Watson defeated two Jeopardy! champions.
And most recently, Google’s AlphaGo defeated one of the world’s best in the ancient Chinese game of Go — a game where you have up to 200 different options to choose from for each move (compared to 40 with chess).
No wonder we’re always battling robots in the movies: we’re battling them in real life too.
And as intelligent machines become more ubiquitous at home and in the workplace, it’s got some people wondering …
“How far will this go?”
“Is my sales or marketing or customer success job at risk?”
“Am I going to be replaced by a robot?”
Why AI is an ally, not a threat
To be clear, machines are taking over some human jobs, especially on the manufacturing side. But this isn’t a new development of the AI movement: it’s been going on for centuries.
Just think about how many scribes slammed down their quills in anger as Gutenberg debuted his printing press. “Welp,” I imagine one of them saying, “We had a good run.”
Technology has always been disruptive. And right now, chatbots are being hailed by some as the next big disruption.
Thanks to messaging platforms like Slack and Facebook Messenger, thousands of chatbots are now accessible to billions of people. And some of those bots seem poised to replace us … at least at first glance.
I think our own chatbot, Driftbot, is the perfect example to look at here.
Driftbot acts like a switchboard of sorts for our messaging app. When a visitor or customer messages us, Driftbot pops up and asks who’d they’d like to talk to, and also asks for an email address just in case they get disconnected.
Now to some, this might seem like Driftbot has taken over a job a human support rep might do. But as someone — a human someone — who actually does customer support here at Drift (everyone on the team has a shift, FYI), that’s not how it works.
Driftbot doesn’t replace you, it helps you.
For example, let’s say you have a bunch of chats come in all at once. Instead of feverishly trying to reply to all of them, Driftbot will automatically step in, say hi, and let people know you’ll be with them shortly. If someone wants to talk to a specific department (e.g. Sales), Driftbot can route them there accordingly.
But it gets better. Based on what people ask in their messages, Driftbot can share links to relevant support docs with them, which means a lot of folks are able to find the answers they’re looking for within seconds of starting a chat.
As you can see, the purpose of Driftbot isn’t to replace customer support, it’s to make our customers’ experiences as amazing as possible. Most of the time that means our customers chat with a human (something we love to do), but sometimes it can mean they get a quick answer from a bot.
Ultimately, we try to have Driftbot do the stuff that bots are good at (re: poring through support docs and matching keywords) and leave the stuff humans are good at (re: talking through a complex product issue) to the humans.
Final thought: Let bots be bots
At Drift, we don’t want our bot to ever become so intelligent that it could replace a human, or trick you into thinking it’s human. We want everyone to know that our bot is a bot.
Why? Because when the lines between robot and human start to blur, things can get pretty creepy.
Ever hear of the “uncanny valley”? It’s a term that describes the revulsion people feel when they see something that looks almost, but not exactly, like a human.
Turns out, we humans prefer robots and other non-living beings that have human-like qualities … but only up to a certain point. When they start seeming too human, it freaks us out.
With robots, you can think about it as a spectrum. On the far left are robots that look unmistakably like robots, and on the far right are humans. In the middle is where things get weird.
While the “uncanny valley” is a visual concept, it still provides an important lesson for chatbot developers: If a bot feels too human, it might end up creeping people out.
At the end of the day, the purpose of chatbots shouldn’t be to create something that can perfectly imitate us (and maybe even someday replace us), it should be to create something can help us.
It shouldn’t be us vs. them, it should be us and them.