Human-AI coevolution
🔹 SES Lab's Journal Club Calendar 🔹
For my first contribution to the series, I’m Momoha Hirose and I’m highlighting:
Pedreschi D. et al. (2025). “Human-AI Coevolution.” Artificial Intelligence, 339, 104244.
The authors reframe recommender-mediated interaction as a society-scale co-evolutionary feedback loop, map the systemic risks it can amplify (e.g., polarisation, inequality, model collapse), and argue that rigorous empirical measurement, together with coordinated scientific, legal, and socio-political action, is essential for keeping the loop aligned with the public good.
This agenda directly informs my own work, which models co-creative decision making between humans and LLMs as distributed Bayesian inference. My broader research interests lie in human–AI symbiotic alignment and collective decision systems, and I hope to advance the authors’ call for “a conscious effort in mathematical modelling to capture feedback-loop mechanisms and their impact on human–AI ecosystems.”
Comments are welcome!
Discussion