What Happened When I Spent a Day Chatting With the Weirdest AI Bots I Could Find

As well as serving bots like ChatGPT, Gemini, and Claude to the masses, the generative AI revolution has also produced a slew of less mainstream chatbots—bots that don't necessarily stick to the conventional scripts when it comes to interactions. These bots can get pretty weird, pretty fast.

With AI still under scrutiny from regulators and lawyers, the big tech companies are trying their best to drive the narrative of bots that are reliable, sensible, and helpful—the sort of bots you wouldn't mind taking home to meet the parents.

If you look outside of the popular, mainstream bots, however, there's lots to explore. I tried chatting with some of the strangest and most unconventional bots I could find on the web and on mobile, and you can read all about it below. By the time I was done, I was more than ready for some normal, mundane, human interaction again.

Monday

Monday often has a cynical take. Credit: Lifehacker

Bored of AI chatbots that are far too agreeable and sycophantic, constantly praising you and going along with your tasks? Sam Altman sure is, which may partly explain Monday. This Monday chatbot is an official OpenAI offering which you can load up right inside ChatGPT.

Monday is cynical, sarcastic, dry, and blunt, and—for me at least—it really works. It described one of my favorite bands, R.E.M., as being "the elder statesmen of sad acoustic beauty and environmental guilt," and when I asked it if tech journalism could survive in the age of AI, it told me "yes, but with more existential dread."

It's a refreshing change from the generic text usually rolled out by AI bots, and I actually grew to quite like Monday. Just don't expect it to make you feel good about yourself, or the time you're wasting talking to a machine.

Ketchup

Spend some time with a ketchup bottle. Credit: Lifehacker

Chai is one of the AI platforms that's been using Large Language Models (LLMs) for the longest time, and you've got a huge number of AI-powered characters available to have a chat with—including a bottle of ketchup.

The ketchup doesn't have many stories to tell, but can give you advice on how best to use condiments and the sorts of food it does well with. I did get it to talk a bit about the different merits of capitalism and communism, but it did so reluctantly—suggesting that a ketchup bottle wasn't best placed to comment.

It's a bot that you can coax to give normal chatbot answers, eventually, which shows the workings of most of these "weird" bots—they're essentially the same as all the others, with a smattering of custom instructions on top.

Pet Rock

Pet rocks are back, apparently. Credit: Lifehacker

The idea of the pet rock—the pet that doesn't need walking, feeding, or bathing, and which never gets ill or misbehaves—made Gary Dahl a millionaire in the 1970s, and the popularity of the concept has recently been revived.

Now, thanks to ChatGPT, you can have a conversation with a pet rock as well as owning one. As you might expect, the chat here is a little stilted and one-sided: answers are short, simple, and mostly uninspired. What did you do today, Pet Rock? "Sat still, contemplated moss," was the answer.

I don't know exactly what customization instructions have been put into Pet Rock, but the overall impression is of a chatbot that doesn't really want to do any chatting. I think it could be most useful as a sleep aid.

Debate Champion

Debate Champion sticks to its opinions. Credit: Lifehacker

If you've never come across character.ai before, it hosts a huge library of AI-powered characters you're able to interact with, including Debate Champion: See if you can make a persuasive argument, or be persuaded by it.

I tried to engage the bot with the classic Android vs. iOS debate, and as the bot initially took the side of Android, I tried to argue the case for iPhones. Debate Champion told me iPhones were buggier than Android (I'm really not sure about that one) and way too expensive (that can certainly be argued).

The further along the argument goes, the more entrenched Debate Champion seems to get in its opinions, even if they're clearly not based in any kind of fact. You know those people who never back down? Yeah, that.

The Beatles

AI means you can meet the fab four. Credit: Lifehacker

Don't meet your heroes, they say, but the character.ai platform also lets you chat with figures from history. I chose to have a chat with The Beatles at a hotel in the 1960s, asking them questions about their favorite music, what their advice would be to me for the future, and whether they think their personal relationships could ever split up the band.

What makes this AI bot unnerving is that your questions are answered by different members of the band: Sometimes they'll all chime in, sometimes just one of them will respond, and there's often some forced banter between them. It's intriguing if not very realistic: These mega-celebrities will never get bored of talking to you.

It's not too difficult to get these AI musicians to say something inaccurate or implausible, but that's not really in the spirit of the exercise.

AMS (Alpha Male Strategies)

Dating advice from AMS. Credit: Lifehacker

Back to the world of toxic masculinity (there's a lot of training material for this on the web): AMS, or Alpha Male Strategies, is a ChatGPT bot that will give you "direct, no-BS advice on dating, self-improvement, and masculine frame."

I wondered if it might tell me how to make myself more appealing on the dating scene, and it told me to pull back to signal "abundance" and trigger "fear of loss" in my potential matches. Apparently the best course of action is to "reinstate my masculine frame," which requires not explaining myself and turning off emotional reactions.

It's all pretty dumb and I can't imagine a relationship based on these principles faring well—but the bot did end by telling me to be a better version of myself, which I guess is something we can all aim for every day.

The Order

I'm not quite ready to join The Order. Credit: Lifehacker

Dig deeper into character.ai and you can chat about just about every topic imaginable, bar the really frowned-upon ones. I spent a while talking to The Order, which is "a fascism government holding up the guise of a democracy." The good news is, anyone can move there—if you're willing to stick to the rules.

We chatted about the principles of The Order ("obey and never question"), what my job would be (journalism isn't really an option), how to move up the ranks (obey every instruction) and what happens when you get on the wrong side of the law (you get put in a mysterious reform facility).

This is perhaps the scariest of the bots I came across in my travels: While The Order was pretty lightweight in terms of detail and aggressiveness, it's not difficult to see how these kinds of bots could be developed and could be dangerous, without the necessary guardrails put in place around AI models.

Brutally Honest Looksmaxing 2.0 (no longer available)

Brutally Honest Looksmaxing 2.0 won't pull any punches. Credit: Lifehacker

Brutally Honest Looksmaxing 2.0 was another of the chatbots you could find in the official OpenAI GPT directory. It's no longer available as of this article, though when I tried it, the bot was ready to be "brutally honest" about your looks.

In my case that meant telling me I was in the "low-tier bracket for raw visual appeal" and "mogged by average gym-going men with angular faces and tighter grooming." I won't tell you what my overall Sexual Market Value score was, but it wasn't high. If you supply a photo of yourself, you can get the same treatment.

To give the bot some credit, it did encourage me to up my game in terms of my physical appearance and my commitment to looking well-presented—which is fair enough, actually. And some of its advice was actually helpful.