Levin Brinkmann
@LevinBrinkmann

New Paper: "Machine Culture"
https://www.nature.com/articles/s41562-023-01742-2

AI systems are increasingly integrated into the human cultural fabric. What can we learn from human cultural evolution to anticipate the impact of intelligent machines on culture (socially transmitted information)? [1/9]

Cultural evolution relies on variation, transmission, and selection. AI systems already operate on all three levels. AlphaGo and DALL-E, are innovating novel strategies and recombinations, humans learn from Chatbots, and recommendation engines curate cultural exposure. [2/9]

We expect future AI systems to increasingly mediate and generate cultural evolution creating “machine culture”. The effect on cultural properties like diversity may range from bad to good. Here are some key aspects; for a more nuanced picture, check out our full paper: [3/9]

Variation: AI can discover solutions beyond human cultural horizons, create complex world models beyond human capacity, and integrate information at an unprecedented scale. This could either enhance or overshadow human creativity. [4/9]

Transmission: AI's role in documenting and preserving culture could improve fidelity in cultural transmission. However, it might also amplify biases or cultural erosion. And yet, complementing AI and human cognition could increase cultural diversity and complexity. [5/9]

Selection: AI shapes our social networks and content creation incentives. It adapts to human feedback with unforeseen outcomes. Machine-generated content is becoming a significant part of cultural production and its curation might become a main driver of cultural evolution. [6/9]

Challenges ahead: Measuring AI's cultural impact is tricky. AI can both diversify and homogenize culture. Balancing these effects is crucial for a healthy symbiosis between humans and machines in culture, but requires ways to quantify effects in the first place. [7/9]

Open questions: How do we navigate trade-offs, for instance between diversity and cohesion in AI-influenced culture? How can we ensure a beneficial coexistence of humans and machines in cultural production? [8/9]

As AI integrates deeper into our cultural fabric, it's essential to anticipate and guide its influence. The framework of cultural evolution can be a useful toolbox to study and anticipate "Machine Culture." [9/9]

Thanks for the wonderful collaboration @ftw_baumann @JFBonnefon @ThomasF_Mueller @amnussberger @MaximeDerex A. Czaplicka @acerbialberto T. L. Griffith @JoHenrich @jzl86 @rlmcelreath @pyoudeyer @iyadrahwan
https://pbs.twimg.com/media/F_ZBFC6WoAMEr8D.png

Nov 20, 2023, 5:01:19 PM

Tweet by @LevinBrinkmann

Paul Bohm — e/acc ⏩
@paulbohm

Come and Take it! At @JUNGSIKDANG with @rivatez and @ErikVoorhees @CBarraford

https://pbs.twimg.com/media/F_GYxeSXgAAh-g6.jpg

Nov 17, 2023, 1:40:51 AM

Tweet by @paulbohm

Charles Rollet
@CharlesRollet1

BREAKING: Hikvision's latest software offers "ethnic minority" recognition technology, even describing the feature as "isEthnic" in an API guide

Hikvision has claimed for years that it stopped supporting this oppressive technology

https://ipvm.com/reports/hikvision-minority-isecure?code=98j98u98u via @ipvideo

Hikvision's face analytics purport to detect whether someone is an ethnic minority (是否少数民族).

Hikvision promptly deleted the API guide off its website after @ipvideo reached out.
https://pbs.twimg.com/media/F9NU4WfboAAMRE3.jpg

Oct 24, 2023, 1:32:51 PM

Tweet by @CharlesRollet1