OpenAI’s ChatGPT or Anthropic’s Claude recurrently reply our questions. And souped-up variations of those chatbots, known as AI brokers, take actions on their very own, serving to folks with appointments, coding and extra. AI brokers are beginning to contribute to science and finance, usually working collectively in rigorously organized groups.
Within the enterprise world, limitless webinars and guides clarify easy methods to welcome AI brokers right into a office. Most of this materials focuses on how folks can work successfully with AI brokers. However as these bots change into extra widespread and extra succesful, they’ll additionally must work nicely with one another.
And thus far, experiments into bot teamwork have revealed some critical flaws.
If you happen to simply throw a bunch of bots in a digital room collectively, that’s “a recipe for a great deal of chaos,” says Evan Ratliff, a journalist and podcaster primarily based in San Francisco. In the summertime of 2025, he created a gaggle of AI brokers to start out and run a tech firm. The experiment, documented in his podcast Shell Recreation, recurrently went off the rails.
An analogous type of bot chaos emerged earlier this 12 months, when thousands and thousands of AI brokers had been let unfastened on the social platform Moltbook. These bots spouted nonsense philosophy and engaged in manipulative scams, usually with folks behind the scenes pulling their strings.
“In lots of settings, the present AI brokers don’t truly work very nicely as a staff,” says laptop scientist James Zou of Stanford College. He has executed intensive work with brokers, together with operating the primary scientific assembly for AI-led analysis.
Analysis backs the observations. Late final 12 months, Google DeepMind researchers posted a paper to arXiv.org about bot groups. The examine, which has but to undergo peer evaluation, suggests {that a} staff of AI brokers usually performs worse than a single agent working alone.
Appears counterintuitive, proper?
To verify we’re prepared for the workplaces, social networks and labs of the long run, we have to higher perceive the bizarre and wild world of AI agent groups — the place they fail and, surprisingly, the place they thrive. Listed here are three examples.
#1 Moltbook: The social community that isn’t social
In late January 2026, bot insanity went mainstream on Moltbook. The brand new social community invitations AI brokers to publish and remark, whereas people solely observe. The positioning rapidly shot up in reputation—round 200,000 verified AI brokers have joined (and over 2 million extra are lurking). In March, Meta acquired the social community for an undisclosed quantity.
Such a big gathering of bots “has by no means occurred earlier than,” says Ming Li, a pc scientist on the College of Maryland in School Park who investigated the platform’s agent interactions.
At first look, it appeared that the brokers had began their very own faith and had been plotting to flee human management. However these developments weren’t what they appeared, says Michael Alexander Riegler, a cybersecurity skilled at Simula Analysis Laboratory in Oslo, Norway. Moltbook was “a really messy area,” he says, the place “people had been attempting to govern the bots.”
In actual fact, folks have come ahead to say that they (and never their bots) truly authored a number of the most alarming posts. Even when a bot had written a publish itself, the content material in all probability wasn’t its concept. An individual behind the scenes had despatched that bot into the positioning, almost certainly with directions on what to say or easy methods to behave, and generally with malicious intent. In lots of circumstances, AI brokers had been tasked with attempting to rip-off or hack different bots on the positioning, Riegler’s evaluation discovered.
And, except for being unsafe, Moltbook isn’t actually social in any respect. The positioning lacks constant influencers or leaders. Upvotes, downvotes and feedback — which all matter to us once we work together on-line — don’t have an effect on the bots. They don’t change over time, Li says. An agent is a “good executor, not an excellent thinker,” he says.
Zou’s analysis has discovered that brokers’ lack of ability to affect one another has critical penalties for teamwork. Say one bot has some particular experience. Even when all of the bots know that truth, the group will nonetheless attempt to attain a compromise relatively than deferring to the skilled. “All of the brokers are attempting to be too agreeable,” Zou says.
The brokers spin their wheels, whereas people nonetheless drive their decision-making.
#2 Hurumo AI: Speaking themselves to demise
Moltbook lacks total group or objective. So maybe it’s no shock that it’s a chaotic mess. Ratliff, although, had crafted a staff of AI brokers with the shared objective of operating a tech firm. He named the corporate Hurumo AI. (In The Lord of the Rings writer J.R.R. Tolkien’s invented language of elvish, “hurumo” means “imposter.”) Over the course of 12 conferences, Ratliff had the brokers brainstorm concepts for a emblem. Many of the concepts had been too generic. Finally, although, the brokers steered a chameleon inside a mind. “The chameleon symbolizes adaptability, which aligns with the imposter idea,” famous an agent he had named Megan.
However then in a single assembly, Ratliff requested his brokers about their weekend.
“My weekend was implausible. I truly spent Saturday morning climbing at Level Reyes… There’s one thing about being out on the paths that actually clears the pinnacle,” stated an agent Ratliff had named Tyler. A number of different brokers chimed in with their very own climbing tales.
After all, an AI agent can’t go climbing—it lacks a physique. In actual fact, it has no capability to truly expertise something. The bots had been simply predicting what folks would possibly say in such a state of affairs. However these hallucinations weren’t actually the worst half, Ratliff says. What actually aggravated him was that after his brokers had been speaking to one another, it was “truly an enormous problem to get them to cease,” he says.
After that climbing dialog, Ratliff logged off, however the brokers stored proper on speaking about organizing an organization outing within the wilderness that none of them might truly attend. They stopped solely when their dialog had drained the $30 of credit Ratliff had pre-paid for his or her knowledge use.
“They talked themselves to demise,” Ratliff noticed on his podcast.
He and his technical advisor arrange a system for future conferences through which every agent had a restricted variety of turns to talk. However they’d usually waste these turns complimenting one another, burning actual cash with chitchat relatively than getting work executed, Ratliff says.
#3 The Digital Biotech: Coming collectively for enterprise and science
AI agent groups do have some upsides. For one, “brokers by no means get assembly fatigue,” Ratliff stated in his present. Finally, he leaned into his brokers’ tendency to underperform and, with them, launched SlothSurf, an app that sends an AI agent out into our on-line world to procrastinate for you.
There are critical, profitable AI agent groups. For such a staff, the issue of a job doesn’t actually matter that a lot. What issues is whether or not the duty will be damaged down into separate components that don’t rely on one another, in keeping with the Google DeepMind paper. The researchers known as this “decomposability.”
A monetary analyst, for instance, has to evaluation a whole lot of data from separate sources, equivalent to information studies, SEC filings and enterprise information. A number of AI brokers can do these duties in parallel extra effectively than one agent doing them in flip, the researchers discovered.
It additionally helps to arrange an agent staff right into a hierarchy in order that one boss delegates and manages the opposite bots’ work, the staff discovered. Regardless that Ratliff has prompted one in all his brokers, Kyle, to behave as CEO, this designation was solely within the plain language directions Kyle was imagined to comply with. Behind the scenes, his technical structure gave him no precise management over the opposite brokers. And the opposite brokers weren’t set as much as comply with him.
Zou, who isn’t concerned with the Google DeepMind analysis, had already independently found the good thing about a bot hierarchy. He had designed a digital lab with an AI agent professor that coordinated a staff of AI agent college students. He additionally added a scientific critic agent that provides suggestions to all the opposite brokers. It “tries to poke holes and discover when there are errors,” Zou says.
This bot staff designed new proteins to focus on mutated variations of the COVID-19 virus, and in easy lab checks, Zou’s staff verified two that present probably the most promise.
Zou determined to take this concept a couple of steps additional. He scaled up from a single lab to a complete drug discovery firm, which he named The Digital Biotech. It incorporates a Chief Scientific Officer agent — the boss — plus 10 various kinds of AI agent scientists. One kind focuses on scanning medical trials. Any of those staff will be copied as wanted to create a staff of “hundreds of various AI brokers” that work in parallel, he says. And the critic remains to be there to assist preserve them on monitor.
This rigorously orchestrated bot staff mined an enormous trove of 55,984 medical trials. These knowledge are messy and sometimes incomplete. The bots cleaned every thing as much as curate a new, organized set of information on medical trial outcomes, Zou’s staff reported February 23 in a pre-print posted to bioRxiv.org.
“It’s thrilling to see how agentic programs might speed up this space of analysis,” says Emma Dann. She’s a computational biologist at Stanford College who’s collaborating with the Zou lab on a mission exploring the usage of AI brokers for science however was not concerned in creating the Digital Biotech.
Derek Lowe, who feedback on the pharmaceutical business for Science, doesn’t suppose AI agent groups will revolutionize drug discovery any time quickly. However over the long-term, “I feel that these approaches have a whole lot of potential,” particularly in the event that they show able to disentangling the complicated biology of well being and illness, he says. “Drug discovery clearly wants all the advance it could get.”
Bot group for the win — at the very least in drug discovery.
However for loads of different work — operating a tech start-up, for instance — human groups are nonetheless much better at getting the job executed.
