Dataset reveals how Reddit communities are adapting to AI

April 7, 2025

The GIST Editors' notes

This text has been reviewed based on Science X's editorial course of and insurance policies. Editors have highlighted the next attributes whereas guaranteeing the content material's credibility:

fact-checked

preprint

trusted supply

proofread

Dataset reveals how Reddit communities are adapting to AI

Researchers at Cornell Tech have launched a dataset extracted from greater than 300,000 public Reddit communities, and a report detailing how Reddit communities are altering their insurance policies to deal with a surge in AI-generated content material.

The group collected metadata and group guidelines from the net communities, referred to as subreddits, throughout two intervals in July 2023 and November 2024. The researchers will current a paper with their findings on the Affiliation of Computing Equipment's CHI convention on Human Elements in Computing Programs being held April 26 to Could 1 in Yokohama, Japan.

One of many researchers' most placing discoveries is the speedy improve in subreddits with guidelines governing AI use. In response to the analysis, the variety of subreddits with AI guidelines greater than doubled in 16 months, from July 2023 to November 2024.

"That is vital as a result of it demonstrates that AI concern is spreading in these communities. It raises the query of whether or not or not the communities have the instruments they should successfully and equitably implement these insurance policies," stated Travis Lloyd, a doctoral pupil at Cornell Tech and one of many researchers who initiated the undertaking in 2023.

The research discovered that AI guidelines are commonest in subreddits targeted on artwork and superstar subjects. These communities usually share visible content material, and their guidelines steadily tackle considerations in regards to the high quality and authenticity of AI-generated photos, audio and video. Bigger subreddits have been additionally considerably extra prone to have these guidelines, reflecting rising considerations about AI amongst communities with bigger person bases.

"This paper makes use of group guidelines to supply a primary view of how our on-line communities are contending with the potential widespread disruption that’s introduced by generative AI," stated co-author Mor Naaman, professor on the Jacobs Technion-Cornell Institute at Cornell Tech, and of knowledge science within the Cornell Ann S. Bowers Faculty of Computing and Data Science. " actions of moderators and rule modifications gave us a novel technique to replicate on how completely different subreddits are impacted and are resisting, or not, using AI of their communities."

As generative AI evolves, the researchers urge platform designers to prioritize the group considerations about high quality and authenticity uncovered within the knowledge. The research additionally highlights the significance of "context-sensitive" platform design selections, which take into account how several types of communities take diverse approaches to regulating AI use.

For instance, the analysis means that bigger communities could also be extra inclined to make use of formal, specific guidelines to take care of content material high quality and govern AI use. In distinction, closer-knit, extra private communities might depend on casual strategies, comparable to social norms and expectations.

"Probably the most profitable platforms can be those who empower communities to develop and implement their very own context-sensitive norms about AI use. An important factor is that platforms don’t take a top-down method that forces a single AI coverage on all communities," Lloyd stated. "Communities want to have the ability to select for themselves whether or not they need to enable the brand new know-how, and platform designers ought to discover new moderation instruments that may assist communities detect using AI."

By making their dataset public, the researchers purpose to allow future research that may additional discover on-line group self-governance and the impression of AI on on-line interactions.

The findings are revealed on the arXiv preprint server.

Extra data: Travis Lloyd et al, AI Guidelines? Characterizing Reddit Group Insurance policies In direction of AI-Generated Content material, arXiv (2024). DOI: 10.48550/arxiv.2410.11698

GitHub: github.com/sTechLab/AIRules

Journal data: arXiv Supplied by Cornell College Quotation: Dataset reveals how Reddit communities are adapting to AI (2025, April 7) retrieved 7 April 2025 from https://techxplore.com/information/2025-04-dataset-reveals-reddit-communities-ai.html This doc is topic to copyright. Aside from any honest dealing for the aim of personal research or analysis, no half could also be reproduced with out the written permission. The content material is offered for data functions solely.

Discover additional

Helper bots in on-line communities diminish human interplay shares