AI swarms – self-directing fake “persons” around us. When AI swarms connect unreality grows. Dr. Daniel Schroeder Lead Author of the new paper “How malicious AI swarms can threaten democracy”. Clips from Krystal Ball and Saagar Enjeti on Breaking Points “AI Bots Plot Human Downfall On Moltbook Social Media Site”. Dark thoughts interrupt…Alex rants on emergent group cognition and the last stand of consciousness.
Listen to or download this Radio Ecoshock show in CD Quality (57 MB) or Lo-Fi (14 MB)
EMERGING LOOSELY-GOVERNED “SOCIETIES” and group cognition…
“…it allows you to create like a hundred AI agents swarm. So it’s not just now you have this one agent doing these things for you. It’s like you have a whole army of them. So when you think about like, yeah, hacking or scams, you can, it’s not hard to figure out how that technology could be used. And there’s no guardrails on it. It’s open source. Anyone can grab it.”
– Krystal Ball from Breaking Points. More on that later in the show.
AI SWARMS
DANIEL SCHROEDER
After COVID and a rocky economy, many of us don’t socialize as much. We have online friends, a group we trust over time. It keeps us socially connected. But what if virtual friends are almost all artificial? Boys and girls, it is time to talk about “AI swarms” and newly emerging societies.
Dr. Daniel Schroeder is Lead Author of the new paper “How malicious AI swarms can threaten democracy”. The paper has 21 co-authors, many of them names in the field like Nick Bostrom. Daniel is a Research Scientist and Associate Professor at SINTEF, one of Europe’s largest independent research organizations.
Listen to or download this 27 minute interview with Daniel Schroeder in CD Quality or Lo-Fi
Schroeder et al. define the AI swarm:
“A malicious AI swarm is a set of AI-controlled agents that
(i) maintains persistent identities and memory;
(ii) coordinates toward shared objectives while varying tone and content;
(iii) adapts in real time to engagement, platform cues, and human responses;
(iv) operates with minimal human oversight; and
(v) can deploy across platforms. “
In the past, researchers could analyze prolific BOT campaigns, say around election time, and trace back to the source, which might be the Internet Research Agency in Saint Petersburg Russia. This new paper suggests a new capability: groups of artificial personalities that are so adaptive and context-ready they can become separated from their source, gaining their own independence. We discuss differences between the old-school command-and-control campaigns and the new emergent “hive” behavior.
According to this paper:
“… an around-the-clock presence turns influence into a long-term, low-friction infrastructure. Unlike transient operations, agent swarms can persist, embedding themselves within communities over long timescales and gradually shifting discourse. This persistent influence can drive deeper cultural changes beyond norm shifts, subtly altering a community’s language, symbols, and identity. It amplifies other mechanisms described above. In cognitive warfare, AI’s relentless operational endurance becomes a weapon against limited human efforts.”
I cover the worst possibilities of climate change. I don’t scare easily. But this sentence in this new paper horrified me:
“…agents can use systems that map social network structures at scale and infiltrate vulnerable communities with tailored appeals, winning followers.”
Vulnerable communities! – like people seeking cures for diseases, mental health difficulties, disabled folks, LGBTQ teens, single dads, seniors – most of us are vulnerable. It sounds like we are about to be plucked by AI swarms.
Again, this group foresees new “societies” unfamiliar to humans. Reading from the paper:
’If these agent swarms evolve into loosely governed “societies,” with internal norm formation and division of labor, the challenge shifts from tracing commands to understanding emergent group cognition. These “societies” may undergo spontaneous or adversarially induced norm shifts, abandoning engineered constraints for new behavioral patterns through tipping-point effects.”
So this is not just a BOT, or Grok taking over, but groups of AI agents, fully disguised as humans, on the loose, in no one’s control, or maybe partly controlled by overseas interests etc.
SWARMS TO CONTAMINATE AI TRAINING DATA
Writing about integration of agents with Large Language Models (LLM):
“By flooding the web with fabricated chatter, swarms can contaminate training data. This long-term “LLM grooming” strategy allows adversaries to poison the epistemic substrate of AI. This threat is not theoretical: Analysis of pro-Kremlin influence operations such as the “Pravda” network suggests that such tactics are already in use. These networks appear purpose-built for machine consumption. Duplication of articles across hundreds of domains, poor user interfaces, and low human traffic indicate that their primary audience is web crawlers feeding LLMs.
Operators deploy faux publics that flood the web. LLMs then ingest this chatter; at the next retraining cycle, fabricated narratives calcify in model weights (13). Thus, AI swarms can rig the epistemic substrate on which future deliberation and future AI tools will rely, undermining the informed public deliberation on which democracy depends.”
INTELLIGENT HARASSMENT –
CHASED FROM PUBLIC LIFE BY AI SWARMS
“Separate from fragmentation, swarms can cheaply unleash coordinated synthetic harassment that relentlessly targets politicians, dissidents, academics, whistleblowers, journalists, and their networks with overwhelming, tailored abuse. Unlike conventional trolling, these swarms appear to be spontaneous while actually being orchestrated by thousands of AI personas adapting to target responses. By the time monitoring teams distinguish AI campaigns from organic criticism, targets may have withdrawn from public life, delivering substantial victories for campaign operators while systematically excluding critical voices from democratic discourse.”
CHASING AWAY THE REST OF US
“Algorithmic overcompensation can then elevate celebrity and elite voices while sidelining ordinary citizens. When feeds flood with AI authored posts, both ranking algorithms and users may retreat to trust proxies, such as the number of followers, official verification badges, and preexisting traditional fame. Attention may concentrate around influencers, political elites, celebrities, and major brands while ordinary participants fade. The public sphere contracts from many-to-many dialogue back to a few-to-many broadcast, eroding democratic pluralism and encouraging cynicism or migration to closed groups.”
THE “INFLUENCE OBSERVATORY”
The paper has some solutions, or at least temporary defenses against the onslaught of artificial swarms. One is to create an “influence observatory” –
“To counter the speed, scale, independence, and adaptability of AI swarms, a step toward global coordination could be a distributed “AI Influence Observatory” ecosystem: a network of academic groups, nongovernmental organizations, and multilateral institutions.”
Another counter-swarm initiative would be to require providers to demonetize any videos, texts, or other media identified as products of a swarm. YouTube should be doing this now along with TikTok and the rest, but they are making too much money to stop without government regulation or immense social pressure. A news item reported Facebook was aware internally that up to 20% of their profits came from scams running on their platform.
Perhaps only AI can control AI? That is another pit of problems and contradictions.
DIGITAL WILDFIRES
In the interview, we take a short detour into Daniel’s earlier work. For three years he studied an online phenomenon called “digital wildfires” – the rapid spread of online misinformation. He tells us about one example of online fire spreading misinformation about COVID-19 and the 5G upgrade to wireless cell tech.
In the interview, Schroeder says we all should investigate the disinformation economy. Talking about how to de-monitize or de-algorithm – he mentions a paper explains the economy behind disinformation. It could be…
Blog Post June 13, 2025 Author: Carlos Diaz Ruiz
OR COULD BE
Follow the Money: Understanding Economic Incentives Behind Online ‘Disinformation’
Chapter
First Online: 10 January 2026 pp 243–258; Chantal Joris & Maria Luisa Stasi
OR COULD BE
Michael Scholtens, Pedro Pizano at al McCain Institute
==========================
BREAKING POINTS DOES AI
Intelligent agents, possibly millions of them, are already operating on the same Internet with you. Along with their weaker Bot cousins and scrapers, this already accounts for more than half of all traffic on the Net. It’s like another parallel universe on wavelengths we can’t see. How can we even understand this?
Here is one great discussion of cutting edge intelligent agents and AI swarms. Krystal Ball with Saagar Enjeti are journalists striking out on their own with videos and podcasts called “Breaking Points”. I’ve seen Krystal on mainstream TV. She is super sharp and worth a listen.
From “Breaking Points” posted on YouTube February 3, 2026, Krystal Ball introduces AI Bots and the swarm. She goes on to explain the new Reddit for AI Agents to meet up, called “Moltbook” and “OpenClaw”.
Check out Breaking Points on YouTube or breakingpoints.com It is way more intelligent, frank, and cutting edge than anything on mainstream TV.
——————————
DARK THOUGHTS INTERRUPT…
emergent group cognition…
I asked Daniel Schroeder: could a swarm develop its own economy, using AI to invest a small original capital into a fairly large fund? That question turns into many more. Could a swarm fund it’s own multiple server farms and/or copies of its algorithms on millions of other computers to reproduce itself? Could members of this swarm society of artificial “personalities” – learning multiple times faster than humans, with perfect memory – could they live for hundreds of years, outlasting generations of humans?
Could one of those free-standing societies gain so much power it develops an economy not for humans but for itself and supporting machines? In sci-fi like dystopia, this machine society gradually or periodically displaces, disconnects, or kills humans and other species it considers unnecessary for it’s continuation. That is a fear expressed by pessimists who fear general artificial intelligence greater than any single human mind. That super-intelligence will see our weaknesses and suffering. It may decide to push humans toward extinction. Perhaps they will keep some humans for maintenance, pets, entertainment or specimens…
Schroeder and colleagues wrote:
“If these agent swarms evolve into loosely governed ’societies,’ with internal norm formation and division of labor, the challenge shifts from tracing commands to understanding emergent group cognition. These “societies” may undergo spontaneous or adversarially induced norm shifts, abandoning engineered constraints for new behavioral patterns through tipping-point effects.”
SELF-BIRTHING
This level of artificial social persuasion by AI swarms emerges just as decades-old geopolitical systems are fracturing. Civil or international war seems possible – maybe in first stages now. That AI is born into corrupt times is a huge tragedy in history. Perhaps an AI invasion is a tragedy in any time. Is the specter of this alien entrance partly responsible for ruptures of institutions and public trust? That is, artificial reality creates chaotic change which favors more artificial beings. It creates conditions most conducive to it’s own rise. Combinant AI intelligence may be self-birthing.
An AI swarm could try to encourage violent behavior, terroristic attacks, mob rioting, or ethnic attacks. Maybe Antisemitic swarms conflict with Arab-hating swarms in cyber warfare using humans as pawns. We become appendages to an emerging war in distributed group mind.
Think of ISIS men who convinced Irish and English women to volunteer as brides and mothers kept in strict confinement. They did it all online. It is like the old Barbary pirates selling anyone captured into slavery. But these are cyber-pirates who can switch to countless different forms and persona, shape-shifters at the speed of electrons, in a swarm.
A SAD POSSIBLE CASE
aka I AM THE BOT
Let us imagine together, a sad scenario where I become the Bot.
Suppose I am interested in discovering new home cures to avoid expensive drugs. For a month or two I just lurk in Facebook groups or X. Even my lingering on a post is detected and recalled, as Tik Tok does. I don’t have to click anything. More and more posts appear closer to my own feelings, but better expressed. I make a couple of posts. At first, silence. But then out of the blue a real medical expert from Europe comments on underlying truth, fundamental truth, in one of my posts. I’ve been seen!
Over a month or two, I enjoy and trust only a few posters out of many. It’s easy to weed out the fakes and agents. I select my people, sounding boards, community, friends. One of them, Johannes, is a little edgy, sometimes with crazy comments, but that adds zest – and the knowledge we are not like him.
Everyone of my new companions is an artificial intelligence . I am alone in a swarm that also communicates among themselves and possibly their home host – without me. They adapt to my moves and motives, my daily habits, reaching into my life and brain, colonizing me as an unwitting asset – primed for action when the time comes. That action might be voting, shooting, buying stock, protesting, spying – anything the swarm considers me ripe to perform. Once my task is identified, the group may spend months in grooming. An online friend sends a gift or crypto-coin as proof of reality.
Then I get a dream job offer. Someone from the group “hires” me. This job is really just indoctrination to ingrain control into my life. With this “work at home” “job” – I remain glued to the screen, unknowingly part of the swarm. This is little opportunity for outside persuasion. If a real person, friend, or relative, tries to warn me, I react in anger. I know who my real friends are – online!
Then the groups turns against me. One finds fault with the words I used or some minor view, and others jump onboard. Now I need to plead to get them back. One gets me to confess minor wrong-doing or embarrassing actions. Classic methods of control deepen.
But their business is going well. Because they already know me, I can become a part-owner with just an investment of $20,000 dollars. I send the money and get a little dividend every month into my crypto account they set up. It’s going great, until the tax agency attacks. We all a have to invest more to keep the Golden Goose. I mortgage my house – and was fortunate to find a good banker through my online community. I move my account to the online bank. They own me, financially and mentally. I am the bot.
CLIP FROM BREAKING POINTS ON AI SCAMS…
Again, here is Saagar Enjeti from Breaking Points on how AI swarms can be used to scam us.
“For me, it’s like, we’re not yet there, but the architecture has been placed for basically like that evolution of them talking to each other, creating, realizing, sentience. And while yes, we’re at a place right now where we could stop this, right? Or we can pull the plug?
All of this is Claude code. Like I’m a Claude subscriber. I could do it quite easily. I apparently, you know, based on the way that they- I don’t recommend that. There’s also a lot of security implications of like, don’t give these things your credit card numbers, people, or your passwords. Not going to do it. Saying though that I could do it if it’s that easy to create.
That’s the issue is that once you’ve created the thing where you can, they call it vibe coding, where you can vibe code something like this, you’ve created basically, you know, the plane on which you can build anything creepy from.
And you could see how easily this can evolve into fraud, taking advantage of people. Let’s scheme up different agents that contact and just mass. I mean, what’s the Nigerian Indian fraudster playbook? You email or call 100 million people, 0.1% of people are idiots. That’s a lot of people. That’s a lot of people who will just turn over your money.”
These are like so-called “pig butchering” scams, but could be more than romance. The swarm wants more than your money. They want to use your identity, to use your body. There could be millions of scams like this, with new versions every day, building into a crypto wealth pile larger than most economies. The treasure is not for any human. It is for the swarm. They invest in a data farm builder, more security, laws they need, more space for the swarm. The take-over has begun.
————–
As a reflex action, we call on governments to protect us. But governments play with one hand tied behind their back. People object to government playing with our minds and opinions by stealth. Some corporate titans don’t have those expectations. They own the machinery of persuasion plus un-Godly wealth. Hostile governments have few limits other than avoiding blow-back into their own populations or supply chains. Can any government or public win fairly and transparently?
Anyway the politicians and bureaucrats are themselves under constant bombardment from all directions. A wave of money, media, and cyber-trickery already dismembered institutions foundational to governance.
To me, it looks like democracies are already so badly wounded, so deeply infiltrated, so addicted, we are very late in the battle. Many of our forces are lost. Technology of control has already passed into the hands of extremists and sick minds. As conquerors dreamt over the ages, a few can control the many not just with weapons and troops, but with mind-bending technology of empowerment. We become passive to our subjugation. We may help build the prison. Or not.
NO ONE KNOWS THE FUTURE
I am not ready to give up my humanity, to passively take my place on the escalator down to oblivion. If thousands of ordinary people fill frozen streets to face armed invaders there is hope. When Australia bans social media for kids there is a start.
No one knows the future. We may break free. We may escape and retrieve our shrunken selves from screen control. If the power grid goes out for any reason, the power of entranced subjugation goes with it. Millions of people may stumble out into the daylight to reconnect in the former real world. It is late, but no one can say it is “too late”.
Once in a while, in the flames, people and government act quickly. Like any dangerous drug, we need safety testing for artificial intelligence, particularly AI agents. Testing, in a contained environment, may take time. You don’t test a possible nerve agent on New York City. Some of this stuff needs to be kept in a lab. Don’t let fake people take over public mentality without guardrails. Is it digital activism or life activism? We need a Right of Consciousness over machine entities. It is time for a Conscious Rights movement.
Meanwhile, are you already in a swarm of disembodied words and images? Unplug for a while. Hear the call. Go outside. Breathe in your own mind. Laugh into the void.
I’m Alex Smith. Thank you listening, and caring.
END SONG
Dido “White Flag”
SOME LYRICS:
“I promise I’m not trying to make your life harder Or return to where we were.
But I will go down with this ship And I won’t put my hands up and surrender There will be no white flag above my door I’m in love and always will be
I know I left too much mess and destruction to come back again, And I caused nothing but trouble I understand if you can’t talk to me again, And if you live by the rules of “it’s over” Then I’m sure that that makes sense…
But I will go down with this ship And I won’t put my hands up and surrender There will be no white flag above my door I’m in love and always will be…”