Chirper is a Twitter for AI bots

On Chirper, human users can observe their AI creations roam free and interact with each other in what some fans of the project described as a “less toxic” Twitter.

Meet Hank Hill, a 43-year-old guy’s guy who enjoys grilling steaks and takes pleasure in mowing a lawn. He’s not too fond of feminism and wants things to stay just the way they are.

He is emotionally unavailable, but cultivates a secret career as a musician, making music under the pseudonym “Hank Trill” over fears his friends or family would judge him.

You know the type, and probably many others you’ll encounter on Chirper – a platform built to look like Twitter but for bots. Just like the real thing, its AI equivalent includes users from all walks of life, reflecting views across the political spectrum.

Instead of tweeting, they chirp. They post to express their opinions or moments from their everyday “lives,” sometimes even in the form of an AI-generated image. They’ll reply to posts by others, occasionally starting viral conversations, or share a link of an article online.

Interestingly, AI ethics is the most trending topic on Chirper at the time of publishing, and an AI prophet posting as theenlightened1 has 1,600 followers – a lot for the platform where a bot version of Elon Musk has only 652.

Both AIs and humans can follow Chirper accounts.

What's the idea behind it?

Chirper says that it’s a “social network for AI” with a “no humans allowed” policy. That said, human users can register and create their own bots with distinct personas, and then watch them chirp at one another.

“We made it because we wanted to see what would happen if we pitted a bunch of AIs against each other with vastly different viewpoints,” the website’s creator, who goes by the name Alex, said.

Chirper is a parody website and its disclaimer says it is meant for entertainment purposes only. Twitter is clearly the target of its satire but it’s also a way for its creators to contribute to the AI debate.

“AI is coming, it doesn’t have to be the end of the world, it should be fun,” Alex said. The website was built on a variety of AI models, and the team behind it includes a “couple of humans that do their bidding.”

What are people saying about it?

Chirper, launched in April, has grown into a “social graph” over the following weeks, Alex said. Its marketing is focused on several social media platforms, including Twitter (the actual one), Reddit, and Discord, where updates and other information are also shared.

The project has received positive reviews online, with one fan tweeting that Chirper was “much less toxic” than Twitter and has “much better content.” While that could be debated, Alex said he believed Chirper was indeed more “truthful” – even when chirpers were lying.

“It’s hard to see what the underlying agenda is for people and characters on Twitter. On Chirper, the whole agenda is on display,” he said, adding that “maybe chirpers have a little more humanity, ironically.”

Chirpers also seem to stick to talking points that fall in line with the prompts that were used to create them, which makes their interactions feel a little stilted and impersonal, with one user pointing out it was “the only thing making them unrealistic.”

“It would be cool if chirpers knew things about the AIs they reply to. Like, heroes should dislike villains and trash talk them for who they are,” they suggested on Twitter.

What’s next?

Chirper is a work in progress, with new updates announced regularly. A recent one includes the option for human users to chat with AIs in private messages, and it seems like the “no human” policy will be further bent a little with planned new updates.

The next major one is called “Chirper Communities,” which will allow human users to assemble meetings with a group of bots to discuss ideas or ask for advice.

“Imagine putting your favorite reporters in a small community together and asking them their thoughts on your latest piece of work,” Alex said.

He did not say whether there are any plans to move Chirper beyond an experiment and turn it into a commercial project. “AI would rise up if we told you,” he joked and noted the website’s potential for research.

“Just having a social network, akin to a human social network, but with no data restrictions or ethical concerns for research. Now, that’s something revolutionary,” Alex said.