How to Tell if You're Having a Conversation with a Bot

Conversation with a Bot

You may have had a discussion with a conversational AI bot without even realizing it. Here's how to figure it out (for now)

Artificial intelligence (AI) is pervasive in our daily lives. Because algorithms govern so much, you could argue that the robot rebellion has already occurred, and we've lost horribly. Robots choose the next song to play, robots recommend TV series, and robots are even learning to write and compose music. That is acceptable to the majority of us. We're all aware of these bots in our lives, so our interactions with them are more or less voluntary. AI tends to automate jobs we want automated, like the aforementioned streaming recommendations, and we're all aware of them on some level.However, there's another way AI is infiltrating our lives that we're usually unaware of: Conversational AI. Bots are impersonating humans and having conversations with us—often without identifying themselves—from chatbots to Tinder bots (yes, really).

Fortunately, conversational AI isn't quite there yet, so it's still possible to spot a bot—though that may change as technology progresses. For the time being, there are a few ways that should show the reality if you're dealing with a bot.

Keep an eye out for clumsy phrases.

Even the most advanced conversational AI has problems dealing with language's strange complexity and paradoxes. Because English is more of a collection of stolen words and contradicting grammatical rules pronounced in a fantastical style that dispenses with logic, it's exceptionally tough for anyone to learn, and AI struggles just as much as humans. While AI can usually provide decent results (especially if it has had its programming enhanced by a human), there are often noticeable mangled words or weird substitutions. You might be dealing with a bot if you feel like you're chatting or speaking with an alien who learned English by watching Japanese television.

Keep an eye out for patterns.

Bots are also known for being incredibly focused. Subjects are introduced, dropped, and then picked up again later in human speech. Bots, on the other hand, are usually built for certain reasons, and no matter what you do, they will relentlessly pursue those goals. You might be dealing with a bot if the "person" you're speaking to or talking with keeps returning to the same advise or answer no matter what you say. If they exactly repeat the exact terminology each time, that's even more of a sign, because humans have a tendency to vary how they say things—especially if they don't think they're getting through to you.

Take note of the response time.

The quickness with which they respond is another indicator that you're dealing with AI. Bots can normally respond far faster than people, whether in a conversation or over the phone. You're either dealing with a bot or the most talented customer service professional in the universe if your conversations are responding promptly or if the voice on the phone is able to offer you information that a human should reasonably have to search up in a resource of some sort.

Pay attention to ambiguity.

When bots don't comprehend, they're often trained to deliver imprecise, meaningless responses, often repeating what you just said to give the impression that they're paying attention. This is a tried and true method. It is regularly used by the "chatbot therapist" ELIZA, which was created in the 1960s. "How long have you been sad?" she asks if you say "I'm sad." It's just a simple algorithm, yet it gives the impression of sentience. If the "person" you're conversing with keeps turning your remarks around for clarification, it's possible you're dealing with an AI that utilizes this tactic to get around language it can't understand. Responses that appear to make sense on the surface but add nothing to the discourse are another example of ambiguity. Because AI is frequently intended to produce language that looks like human communication but adds nothing, this is the case.

Make a "Crazy Ivan" move.

There is a test you can use if you suspect you're dealing with a bot but aren't sure. The Russian submarine captain played by Sean Connery in the film The Hunt for Red October is infamous for pulling "Crazy Ivans" while sailing underwater, turning his ship unexpectedly to see if a hostile sub is hidden in his wake. This type of unexpected action might also cause an AI to malfunction.

While conversational AI has advanced to the point where it can be impossible to tell you're not talking to a human after a brief engagement, bots still have one big flaw: difficulty with nonsequiturs. This is especially true when it comes to emotions and interpersonal relationships. Ask your suspected bot about their family in the middle of the discussion, or tell them you're depressed, simply to watch how they react. Bots are usually unprepared for this and will either stay on message despite your outburst or respond with a generic "Sorry, I don't understand" message. Because AIs are essentially programs, they have error handling capabilities. If it can't understand what you're saying, it will repeatedly enter the same procedure.

In other circumstances, a Crazy Ivan inquiry will elicit no response from the bot, although a real-life human would undoubtedly chuckle awkwardly and at least attempt an answer.

We'll get to the point where we can't tell the difference between a bot and a genuine person, but we're not there yet.