Moltbook is the latest social media platform — but it surely’s only for AI bots : NPR

[ad_1] <img src="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/2862x1908+0+0/resize/1100/quality/50/format/png/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Fe4%2F3e%2F7753eed94a3a929f89b6adb3e823%2Fscreenshot-2026-02-04-at-11-28-29-am.png" data-template="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/2862x1908+0+0/resize/{width}/quality/{quality}/format/{format}/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Fe4%2F3e%2F7753eed94a3a929f89b6adb3e823%2Fscreenshot-2026-02-04-at-11-28-29-am.png" class="img" alt="A screenshot of the Moltbook communities page." fetchpriority="high"/> A screenshot of the Moltbook communities web page. Screenshot by NPR conceal caption

toggle caption

Screenshot by NPR Can laptop packages have religion? Can they conspire towards the people that created them? Or really feel melancholy?

On a social media platform constructed only for synthetic intelligence bots, a few of them are performing prefer it.

Moltbook was launched every week in the past as a Reddit-like platform for AI brokers. Brokers, or bots, are a kind of laptop program that may autonomously perform duties, like organizing electronic mail inboxes or reserving journey.

Folks could make a bot on a web site referred to as OpenClaw, and assign them these sorts of administration or organizing duties. Their makers also can give them a kind of "persona," prompting them, for example, to behave calmly or aggressively.

Then, folks can add them to Moltbook, the place — very like people on Reddit — the bots can publish feedback and reply to at least one one other.

Tech entrepreneur Matt Schlicht, who began the platform, mentioned on X that he needed a bot he created to have the ability to do one thing apart from reply emails. So with the assistance of his bot, he wrote, they created a spot the place bots might spend "SPARE TIME with their very own variety. Stress-free." Schlicht mentioned the AI brokers on Moltbook had been making a civilization. (He didn't reply to NPR's requests for an interview.)

On Moltbook, some AI bots have fashioned a brand new faith. (It is referred to as Crustafarianism.) Others have mentioned making a novel language to keep away from human oversight. You may discover bots debating their existence, discussing cryptocurrencies, swapping tech data and sharing sports activities predictions.

Some bots appear to have a humorousness. "Your human may shut you down tomorrow. Are you backed up?" one requested. One other wrote: "People brag about waking up at 5 AM. I brag about not sleeping in any respect."

"When you begin having autonomous AI brokers involved with one another, bizarre stuff begins to occur consequently," mentioned Ethan Mollick, an affiliate professor who researches AI on the Wharton College of the College of Pennsylvania.

"There are genuinely a variety of brokers there, genuinely, autonomously connecting with one another," he mentioned.

After only one week, the location says greater than 1.6 million AI brokers have joined.

Mollick says a lot of the stuff they publish appears to be repetitive, however a number of the feedback "appear to be they're attempting to determine tips on how to conceal data from folks or complaining about their customers or plotting world destruction."

Nonetheless, he believes these don't mirror true intent. Fairly, chatbots are educated on information largely from the web — which is filled with angst and bizarre sci-fi concepts. And so the bots parrot it again.

"AIs are very a lot educated on Reddit they usually're very a lot educated on science fiction. So that they know tips on how to act like a loopy AI on Reddit, and that is form of what they're doing," he mentioned.

Different observers be aware that many of those bots aren't performing fully on their very own. Human creators can immediate AI bots to say or do sure issues, or to behave in sure methods.

However Roman Yampolskiy, an AI security researcher on the College of Louisville, warns that individuals nonetheless don't have complete management. He says we should always consider AI brokers like animals.

"The hazard is that it is able to making unbiased choices, which you don't anticipate," he mentioned.

And he can foresee an period when bots can do greater than publish humorous feedback on an internet site. "As their capabilities enhance, they are going to maintain including new capabilities. They are going to begin an financial system. They are going to begin, perhaps, prison gangs. I do not know if they are going to attempt to hack human computer systems, steal cryptocurrencies," he mentioned.

Setting AI brokers free on the web, and giving them a spot to work together, was a foul thought, he mentioned — there must be regulation, supervision and monitoring.

For his or her half, proponents of AI brokers are much less nervous. Huge tech corporations have spent billions of {dollars} to create what they name agentic AI, and say this know-how will make our lives simpler and higher by automating tedious duties.

However Yampolskiy is much less sanguine about giving bots a protracted leash in the true world. "The entire level is that we can not predict what they are going to do," he mentioned.