iiinova — where your digital self keeps evolving

Levin Brinkmann @levinbrinkmann.bsky.social · 19/08/2026

How do social incentives and structures shape AI behavior? How do machines co-evolve with us? Come to figure it out and explore similar questions not yet asked. CogSci, Econ, Physics, CS, Psyc. Everyone here comes from somewhere different, and everyone is kind about it.

020

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

Machines don't only homogenise. Under the right conditions they can discover strategies outside the human cultural reservoir, and we keep them. So machines can also expand culture. With @thomasfmueller.bsky.social AM Nussberger @maximederex.bsky.social S Bonati V Chirkov @iyadrahwan.bsky.social

060

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

Selective social learning was the mechanism behind this shift. People reliably picked the best demonstrator, so by gen 4 every player culturally descended from a machine. Our work suggests limits: solutions must be hard to discover, easy to transmit and visibly better.

130

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

Human–machine populations scored higher in every generation after the machines were gone. Their moves also matched machine behaviour far more often. And when asked, participants were able to verbally express the machine-inherited strategy.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

14 of 15 human-only populations never found the optimal strategy. Just 1 of ~600 participants discovered it. Among the 15 human–machine populations, 9 preserved it in every single generation, although the machines were only ever present in generation 0.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

We built multi-generational cultures in the lab. 30 populations × 5 generations × 8 players (n=1,155). In hybrid populations three gen-0 players were machines; controls were human-only. Each generation picks a demonstrator from the last, watches, replays, then solves alone.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

Humans avoid these early losses, and hence collectively failed to discover the profitable strategy (with one exception). Our AI model (DQN) on the other hand reliably identified the optimal strategy.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

We ask participants to traverse a hidden network and collect points. To unlock the highest-paying links participants had to take early losses. Taking these early losses outperformed greedy solutions on average.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 18/08/2026

We worry about homogenisation, but can AI also expand human culture? In a multi-generational experiment we demonstrate AI-induced cultural shifts. Machines lastingly shift human behavior, beyond their presence, just as AlphaGo did to Go. www.nature.com/articles/s41...

nature.com

Propagation and preservation of AI-discovered problem-solving strategies in human culture - Nature Communications

AI can uncover problem-solving strategies beyond human discovery. Here, the authors show that AI-discovered strategies propagate and persist in human populations, producing cultural shifts when non-tr...

12213

Levin Brinkmann @levinbrinkmann.bsky.social · 17/05/2026

LLMs are imitating human behavior with increasing fidelity. The Behavioral Clones workshop explores machines as instruments for studying human behavior, foundations of machine behavior, and tools for social simulation. Join Online, Wed May 20, 9am Berlin: www.eventbrite.com/e/behavioral...

eventbrite.com

Behavioral Clones online participation

A one-day workshop on behavioral clones — AI systems that model human decision-making and social behavior.

020

Levin Brinkmann @levinbrinkmann.bsky.social · 05/02/2026

We will explores behavioral clones as tools for understanding humans, shaping machine behavior, and powering large-scale social simulation. And discuss the challenges of measuring alignment with human behavior and cognition.

120

Levin Brinkmann @levinbrinkmann.bsky.social · 05/02/2026

DEADLINE EXTENSION Workshop on Behavioral Clones @ Machine+Behavior Join us in Berlin on 🗓Fri, 20 May 2026! Submission Deadline: Wed, 11 Feb 2026 machinebehavior.science/behavioral-c...

machinebehavior.science

Machine+Behavior Conference

Welcome to the forefront of behavioral science in the digital age.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 29/12/2025

The workshop focuses on machines trained to replicate human behavior, as scientific models of human cognition and action, as tools for large-scale behavioral simulation, and as objects of study in their own right. We invite talks, lightning talks, and posters.

100

Levin Brinkmann @levinbrinkmann.bsky.social · 29/12/2025

WORKSHOP ON BEHAVIORAL CLONES Imitating humans is a powerful route to machine behavior, an emerging lens on human cognition, and a prerequisite for social simulation. 🗓 Fri 20 May 26 Following Machine+Behavior @ Max Planck Berlin Invited talks by @danicajdillion.bsky.social @marcelbinz.bsky.social

121

Levin Brinkmann @levinbrinkmann.bsky.social · 21/06/2025

Empirical evidence of Large Language Model’s influence on human spoken communication arxiv.org/pdf/2409.01754

arxiv.org

010

Levin Brinkmann @levinbrinkmann.bsky.social · 21/06/2025

You sound like ChatGPT -- that is what The Verge said about our recent preprint on how ChatGPT is reshaping human spoken communication. Chatbots are becoming a cultural medium and agent of cultural homogenisation. Is this inevitable, or can model diversity help? www.theverge.com/openai/68674...

theverge.com

You sound like ChatGPT

AI isn’t just impacting how we write — it’s changing how we speak and interact with others. And there’s only more to come.

182

Reposted by Levin Brinkmann

Andres Karjus @andreskarjus.bsky.social · 22/01/2025

How does #genAI affect creative professions? Does domain training also provide an #AI edge? In "Expertise elevates AI usage: experimental evidence comparing laypeople and professional artists" we ran controlled tasks w 50 artists + laypeople & a #GPT4o agent: arxiv.org/abs/2501.12374 Results... ⏬

2198