Help! I Think I’m Dating a Robot!
In this age of AI, you may find yourself wondering at what point the line between human and AI becomes indistinguishable - and you wouldn’t be the only one! A few years ago, I had the pleasure of taking a class all about cyborgs, where I became increasingly uncomfortable imagining the many ways our lives are already intertwined in the existence of mechanical enhancements and artificial intelligence. My biggest takeaway from that class was that none of us should be surprised if we one day find ourselves walking alongside AI-powered humanoids who speak and act the same way as the rest of us, but here at Blink we have one major concern: what does this mean for dating?!
AI is popping up everywhere, and popular uses range from planning trip itineraries to preparing blog posts (this one was written by a human!) to writing dating app DMs. So how can you know that the person you’re talking to is actually investing their energy in you and not just passing off your conversation to a robot? Or scarier yet, how do you know they’re a real person?!
Don’t fret, we got you.
A Quick Poll
As an experiment, I asked four of my coworkers if they thought they would be able to tell the difference between a real person and AI. One responded “I honestly don’t think I could” while another confidently said “I could definitely tell the difference”. The other two were uncertain. Overall, it was an even split that left me wondering whether people can distinguish messages written by a human from those written by AI, and how confident we are in our ability to do so.
Spotting the Difference: 6 Quick Tricks
In the noble name of love, I scoured the Internet for ways to identify the difference between human-written and AI-generated content. Drumroll please…
Trick #1: Pay Attention to Tone & Voice
Not to sound like your high school English teacher, but tone and voice are key to spotting the differences between something written by an actual person and something written by artificial intelligence (no guarantees that this works though – some people do have the social skills of a robot).
Snippets written by AI tend to sound more professional and formal, often spelling out contractions (“I am” instead of “I’m”) and limiting the amount of emotion and passion present in the writing. On the other hand, humans - particularly those communicating with people they have some kind of vested interest in - tend to be a bit more relaxed with their word choice and demeanor. This trick alone is not certain to tell the difference; some people may prefer to kick off the conversation with a more formal “Hey, how are you?” rather than a casual “What’s up?” and both would be valid lines from both a real person and an AI chatbot, but if you find the “person” you’re talking to is consistently maintaining a formal tone and sense of professionalism similar to the emails you send your boss, you may be talking to a robot.
Trick #2: It’s All About the Personality
Look, I don’t know enough about AI to argue about whether or not it can have a personality, but if your chat buddy never reveals any sense of humor and always sounds about as interesting as a calculus textbook, it might be time to question if they’re an actual person or not. AI-written blurbs tend to lack the personality and humor that most of us knowingly and unknowingly communicate with, causing the writing to feel like it’s laying out the facts but not actually interpreting anything. A big tell here is if the other person is able to communicate empathetically - for example, if you say that you’re stressed and your chat partner can only respond with a generic statement along the lines of “How can I help?” or over-detailed instructions to find a solution, the lack of empathy and/or robotic response may be a key signal that you’re not talking to a human (or at least one capable of understanding empathy!).
ChatGPT can be great for offering helpful, objective advice… but not so great at empathizing.
Trick #3: Slang & Emojis
You may also notice that even if the writing has some kind of personality, it may come across as trying too hard. Humans generally know how to balance the contents of their writing to sound, well, human-like, but AI hasn’t quite mastered this (yet). Over-utilization of slang could also be a sign that the “person” you’re talking to isn’t real.
Another possible tell is the use, lack of use, or misuse of emojis. I personally have never seen AI chatbots use emojis, but based on how they get recommended when I type in keyboard shortcuts, I can only imagine how off-putting the emojis might come across if a chatbot did try to use them. For example, when I type in “:)”, I get the recommended 🙂 (“slightly happy”) face, which to me comes across way creepier and less friendly than something like 😊. In my unprofessional opinion, if you’re regularly and unironically receiving the 🙂 emoji from your chat buddy, it’s time to dip.
Trick #4: Ask Your Gen-Z Friend
If you don’t already have a Gen-Z friend, I highly recommend finding one. Even as a Gen-Z’er myself (although I really consider myself part of a separate generation, having been born in one of those in-between years from 1997 to 2003, but I digress), I often find that having younger Gen-Z friends helps a ton in understanding the “modern” world. A recent study found that Gen-Z’ers were statistically better at distinguishing AI-generated images from human-created images (Rajnerowicz, 2023). While the findings from this survey were specific to distinguishing images, it wouldn’t be far-fetched if the same thing were true with text. Even within the findings from my quick poll mentioned earlier, the two coworkers who are Gen-Z were more confident in their ability to tell the difference between human-generated and AI-generated content (at least in some circumstances), whereas the other two (both technically Millennials) were less confident. I guess growing up surrounded by screens has some perks…
Trick #5: No Pics, No Proof
I would like to preface this section with the fact that I don’t think asking for photos from a stranger on the Internet is the best idea, but if you seriously can’t tell and are afraid to ask, this may come in handy. Granted, they could send you an AI-generated image, but it’s hard to get multiple images of the same person from AI, so make sure to ask for a few pics in different contexts (or hit up your Gen-Z friend to take a look). If they do send an AI-generated image, that’s a pretty huge red flag. What do you have to hide, other than the fact that you have no face? 👀
Trick #6: Pin-AI-occhio Could Never
One thing Pinocchio couldn’t do without a tell: lie. One thing AI may not be capable of intentionally doing, at least not without eventually giving it away: lie.
At the end of the day, if you’ve tried everything you can think of to figure out if your chat buddy is a real person or not, maybe it’s time to just ask the question. I tried it with the MyAI feature on Snapchat, which was brutally honest about its humanity.
Snapchat’s MyAI feature gets real about it’s humanity and ability to lie.
I will say that the Internet is full of articles regarding AI’s capacity to lie, but everything I found that said this was in regards to chatbots making up facts, not entire life stories. If you ask it for statistical data, a code snippet, or quotes from a book, it very well may make up the information it gives you… but will nonetheless present the information as cold, hard facts. If you call it out for being wrong or giving you fake quotes, it will give a quick apology before spitting more information out, which may also be made up or incorrect. It isn’t doing so to intentionally lie – rather, it’s using the information it has and filling in the gaps for what it’s missing so that it can give you what you’ve asked for.
With all that said, who actually knows what AI is fully capable of?! Just know that once you ask the question of whether or not your chat buddy is a real human, you can’t really go back – so hopefully if they are a real person, they find the exchange funny rather than insulting.
And, if they refuse to ever talk to you again, you can always download Blink and try again 🤪 – at least you’ll know the person on the other end of the call is real (because AI hasn’t gotten good enough to replace us on calls… yet 😅).
