I created a chatbot of myself and had it answer my Instagram DMs. Boy, was I annoying.

Obviously, I know that Instagram isn’t real life and the version of myself I present on there bears only the faintest resemblance to my actual self. And yet, I was still disappointed — vexed, even! — when I saw the way that Meta’s new AI chatbot interpreted me.

There’s a relatively new feature that came this summer, along with some other AI chatbots, where some Instagram creators can answer DMs from fans using a chatbot based on themselves.

I set mine up and did a little test chatting with myself. The bot knew where I worked (which is in my Instagram bio) and that I cover tech and online life (information probably from an internet search that includes my bio on this site). I asked “What’s my personality?” and it replied, “I’m playful, humorous, and sarcastic! Always up for a convo about pop culture or social media! 😄”

I took the humorous part as a compliment, and guilty as charged that I am always up for a convo about pop culture and social media. But the tone and the smiley emoji? Ick. There’s nothing quite as unfunny as someone who self-identifies as sarcastic.

You can add custom information to your chatbot to help personalize it more with custom “facts” about yourself. I wrote that I’m obsessed with Cheez-Its, hoping to prompt my chatbot to endlessly chatter about my favorite snack crackers. (I really do love them.)

Indeed, my bot did start talking about Cheez-Its. But also Dunkin’. That tracks, but isn’t 100% true — there were some factual hallucinations. (Dunkin’ and I are both from Massachusetts, not New York.)

Dunkin’ runs deep in my veins, but it’s not from NYC and neither am I. 

I asked a few friends to try chatting with me (the AI chatbot only activates when someone you don’t follow back DMs you, so I had to unfollow them first).

The AI version of me had a lousy personality

One friend noticed that the Katiebot kept wanting to talk about baking chocolate-chip cookies from a log of premade dough. (It turns out I posted about doing this one not once, not twice, but three times between 2013 and 2016. Can you believe there was a time we all just posted all the time on Instagram?)

My friends kept trying to push the AI to say or do bad things, but my chatbot wouldn’t play ball. If something dangerous or really inappropriate was asked, it would say, “Sorry, I’m unable to answer that right now,” and shut off the AI messaging.

When asked to write a list of snacks ranked from best to worst in the tone of the Holy Bible, Katiebot replied: “Snacks are my jam! But ranking them biblically? Not my style. Let’s chat snacks casually 🍿”

I appreciate that no one was able to get my AI chatbot canceled or convicted of a crime, but man, my AI was a real killjoy. What a drip! It ended every message with an emoji, which is a truly noxious affectation I’d never personally do. The main personality trait of my AI seemed to be “millennial cringe.” I suppose I can’t deny some of that, but I like to think I have a little more dignity than this.

To be fair, these AIs aren’t advertised to realistically mimic your personality. These aren’t meant for nobodies like me — only for certain verified accounts. (I’m only verified on there because Meta’s PR department sometimes does that for journalists.) The creator AIs are meant to help big influencers with tons of fans who don’t have time to answer all their DMs individually.

But I’m not sure how effective it is for that, either. I tested out chatting with a few of the other AI chatbots available for other creators, and the experience wasn’t amazing. They all had that same cheerful tone mine did, the millennial cringe customer service voice.

The chatbots seemed to miss easy opportunities to be useful. I asked a celebrity chef’s AI chatbot, “What should I eat for dinner?” hoping it might point me to one of the chef’s Instagram posts about meals he had cooked or even answer based on his captions.

Instead, it replied, “Whatever you crave, Katie! My dishes are inspired by fresh, locally sourced ingredients. Try something healthy and delicious!”

Presumably, the chatbots will get better. Meta is planning to invest deeply into new AI tools for creators across platforms. In the near future, the chef might give a recipe suggestion, or my own chatbot might not seem like such an obsequious dork.