How to Tell if You’re Chatting with a Bot

Artificial intelligence (AI) is invading every aspect of our lives. So much is ruled by algorithms you could make the case that the robot uprising has already occurred and we lost, and lost badly. Robots decide what song plays next, robots recommend TV shows, and robots are even getting pretty good at writing and creating music. Most of us are fine with that. AI tends to automate tasks we want automated, like the aforementioned streaming recommendations, and on some level we’re all aware of these bots in our lives, so our interactions with them are more or less voluntary. But there’s another way that AI is invading our lives, and we’re not usually even aware of it: Conversational AI. From chatbots to Tinder bots (yes, really), bots are pretending to be human and having conversations with us — often without identifying themselves.

Luckily, conversational AI hasn’t yet reached perfection, and it’s possible to detect a bot — though that may change soon enough as technology advances. For the time being, if you want to know if you’re dealing with a bot, there are a few strategies that should reveal the truth.

Watch out for clunky phrasing

Even the best conversational AI often has trouble with the weird complexities and contradictions of language. English is especially difficult for anyone to learn because it’s less a language than a collection of stolen words and contradictory grammatical rules pronounced in a fanciful way that dispenses with logic, and AI struggles as much as human beings. While AI can usually pull off pretty decent results (especially if its scripting has been augmented by a human being), there are often telltale garbled sentences that simply make no sense, or strange substitutions. If you feel like you’re chatting or speaking with an alien who learned English by watching Japanese television, you may be dealing with a bot.

Look for repetition

Bots also tend to be extremely single-minded. Human conversation tends to be fluid — subjects are introduced, dropped, and then picked up again later. But bots are usually constructed for specific purposes, and they will doggedly pursue those purposes no matter what you do. If you notice that the “person” you’re speaking to or chatting with keeps returning to the same recommendation or solution no matter what you say, you might be dealing with a bot. If they literally repeat the precise phrasing each time, that’s an even stronger indication, because humans tend to change how they phrase things — especially if they sense they’re not getting through to you.

Note the response speed

Another sign that you’re dealing with AI is the speed of their responses. Whether in a chat or on the phone, bots can usually generate responses much faster than humans. If the chats are coming back to you instantaneously, or if the voice on the phone is able to instantly give you information that a human should reasonably have to look up in a resource of some sort, you’re either dealing with a bot or the most talented customer service rep in the universe.

Pay attention to vagueness

Bots are often programmed to offer vague, meaningless responses when they don’t understand, often repeating what you just said in order to give the illusion of paying attention. This is an old trick. The “chatbot therapist” ELIZA, developed in the 1960s, uses it constantly. If you tell her, “I’m sad,” she responds “How long have you been sad?” It’s a simple algorithmic construction, but it offers the illusion of sentience. If the “person” you’re chatting with constantly turns your statements around for clarification you might be dealing with an AI that uses this trick to get around sentences it can’t parse easily.

Another aspect of vagueness is responses that superficially make sense but add nothing to the conversation. That’s because AI is usually programmed to produce language that resembles real communication but actually adds nothing.

Pull a “Crazy Ivan”

If you suspect you’re dealing with a bot but you’re not sure, there’s a test you can try. In the movie The Hunt for Red October, the Russian submarine captain played by Sean Connery is known for pulling “Crazy Ivans” while sailing underwater — suddenly turning his boar to see if an enemy sub is hiding in his wake. This kind of surprise move can disrupt an AI as well.

While conversational AI has become extremely sophisticated and it can be difficult to tell from a brief interaction that you’re not talking to a human being, bots still have one major weakness: Difficulty with non sequiturs. This is especially true with emotions and human relationships. In the middle of the conversation, ask your suspected bot about their family, or tell them you’re feeling depressed, just to see the reaction. Bots typically won’t be prepared for this and will either stay doggedly on message, ignoring your outburst, or offer a generic “Sorry, I don’t understand” kind of message. That’s because AIs are ultimately programs, and thus have what’s known as error handling. If it can’t process what you’re saying, it will drop into the same subroutine over and over again.

In some cases, the bot won’t respond at all to a Crazy Ivan question — whereas a real-live human will probably laugh nervously and at least make some kind of effort at a response.

The time will come when we can’t tell the difference between a bot and a real human being — but we’re not quite there yet.