29 May 2017
How to tell if you are talking to a chatbot
Chatbots are everywhere these days…
It’s no secret that bots are taking over messaging platforms, especially on behalf of companies. In fact, many initial interactions between consumers and companies are already bot-driven. Some are better than others, but all are getting smarter about interacting with their human counterparts in a natural manner.
Simply put, bots are programs designed to respond in a certain way depending on keywords they receive. Many companies use these bots to interact with consumers, particularly when it comes to their customer service functions and social media sites. These chatbots help them respond to more users at a faster and cheaper rate than employing human customer service representatives. If you haven’t already, you’ll soon notice a lot of your favourite brands rolling out these “virtual responders”.
On the other end of the spectrum, bots can also be found lurking in messaging apps and community forums. Such bots aren’t always upfront about being a computer program rather than a real person. Chances are that if you’re on Tinder you’ve already come across a bot (or a hundred). It may even be the case that one or two of them fooled you for a while before you realised that you were chatting with a computer program.
With the linguistic lines being so blurred between real people and chatbots, how can you work out if you’re talking to a chatbot?
While it’s becoming more difficult to tell, or at least tell right away, there are still a slew of signs that can reveal the truth, if you know what to look for…
How to tell if you are talking to a chatbot
…they respond lightning fast.
Even when peppering their conversation with LOLs and emojis, real people tend to take longer than a split second to answer messages. If you get a super fast response, especially when it’s more than a simple yes or no response, you can bet you’re chatting with a bot.
…they don’t speak naturally.
‘Unnatural’ speech can encompass a number of different things. If the responses you get are always in full, complete sentences, be suspicious. Most people tend to message in fragmented sentences and thoughts when they’re chatting, so if the ‘person’ you’re talking to uses full and formal sentences ‘they’ may be a bot. Alternatively, some bots have been programmed to go against robotic-sounding formality to come across overly familiar which can also sound unnatural. If you receive an inordinate amount of acronyms, messaging slang and emojis, those should all also be red flags.
…they repeat the same message.
Got the same answer to two different messages? It’s a bot. While bots are programmed to respond to a range of triggers and situations, it’s impossible to program them with an infinite amount of responses. Real people, on the other hand, would never respond with the exact same message in the same conversation.
…they say they don’t speak your language well.
Of course, if you’re chatting with someone in their second language you’ll probably notice more errors than when you speak with a native English speaker. This line is used by bots, particularly on dating apps, as a way to establish a level of leniency to let language errors go unnoticed and have you continue thinking you’re talking to a real person.
…they use strange syntax.
Okay, people often make little mistakes when they message, like a missing comma here and extra space there, but if you repeatedly notice two spaces between sentences, strange punctuation or indentations, chances are there’s a glitch in the computer you’re talking to.
…they mention a product or service you’re not using.
If your chat partner mentions a product or service seemingly out of nowhere, it’s most likely a disguised sales pitch.
…they send a link unasked.
This will often happen on dating apps or forums. You’re chatting and going about your business then – BAM – suddenly you get hit with a link to buy male enhancements.
…they ask for personal or financial information.
No one should be asking for your social security or credit card number through a messaging system.
Run away immediately.
How to catch a chatbot
If you’re having a chat and things start to sound a little funny, you’ll start questioning ‘who’ it is you’re talking to. Want to try and catch a sneaky chatbot in action? Try one of these simple tricks to find out if you’re chatting with a bot.
Ask “them” about themselves
What are their hopes and dreams? If it’s a chatbot, the program will eventually trip up trying to keep proving that it’s ‘real’.
Just say “Um”
A lot of chatbots get confused when you start using onomatopoeia fillers like “um”, “uh-huh” and “hmm”. If you use them in your messaging and get a general or non-sequential response back you’re likely talking to a bot.
Sarcasm can be hard to pin down via text, even if you’re chatting to a friend. Bots have a really hard time with it and will most likely respond to your sarcasm literally.
Bring up current events
Ask about something major that just happened in the world and see what they say. If it’s a bot, it probably won’t be able to keep up with intelligent answers.
Keep asking “Why?”
The same way a parent will start coming up with nonsensical answers to their toddler’s incessant “Why? Why? But why?” questions, a bot will start to get confused and run out of answers if you just keep asking “Why?”
Does it matter?
Chatting online with a chatbot can often actually be quite helpful. They can provide useful troubleshooting information and help solve technical issues, often quicker and more efficiently than a human can.
Chatbots are a useful tool. In the right context, they can help brands solve all kinds of very real customer service problems.
But in the wrong context, chatbots can be annoying at best and dangerous at worst, which is why it is often better to know who (or what) is on the other end of that chat you are having.
Have you ever found yourself chatting with a bot that you at first thought was a real person? What clued you in?