Reddit Deploys AI to Combat AI-Generated Spam

A new study by Pew Research indicates that more than one-third of web pages published since ChatGPT's launch in November 2022 exhibit signs of AI authorship or substantial AI editing. This finding suggests a significant shift in web content creation, with implications for information quality and the future of online content.

The widespread adoption of AI detection tools by educators and publishers is creating distrust due to their dubious accuracy and subjective evaluation methods. These tools, unlike traditional plagiarism checkers, use AI models to guess if text is human-written, leading to potential false positives and concerns about fairness.

AI moderation tools on social media platforms, such as Reddit, can lead to the erroneous removal of valuable user-generated content, undermining the authenticity and archival value of communities. This issue highlights the limitations of relying solely on AI for content moderation, as it can misinterpret legitimate content as spam and delete it without human oversight.

YouTube has updated its Partner Program policies to restrict certain types of AI-generated content from monetization, shifting from "repetitious content" to "inauthentic content." This change aims to ensure creators using AI still demonstrate effort and creativity, impacting how some AI-generated videos can earn revenue on the platform.

Snapchat will no longer recommend wholly AI-generated videos in its Spotlight feed, following similar actions by YouTube, LinkedIn, and Substack. This change addresses concerns about the proliferation of low-quality, repetitive AI-generated content and its potential to erode trust in online information.

Snapchat will no longer recommend entirely AI-generated videos in its Spotlight public feed to prioritize authentic, human-made content. This change aims to reduce low-quality, repetitive AI content, even if creators disclose its AI origin, while still allowing AI tools for content enhancement.

Snapchat announced it will no longer reward fully AI-generated videos in its Spotlight recommendation system, prioritizing human-made content. This change aims to maintain the platform's focus on authentic creativity and personal storytelling, aligning with a broader industry trend of social media platforms adjusting policies regarding AI-generated material.

LinkedIn has added a new 'Seems like AI slop' reporting button for users to flag low-quality, AI-generated content. This initiative aims to address user frustration with the prevalence of such posts and complements LinkedIn's automated defenses and new classifiers for identifying low-quality content.

LinkedIn has added a new reporting option allowing users to flag posts as "Seems like AI slop" and is implementing new classifiers to identify low-quality, AI-generated content. These changes aim to reduce the amount of AI-generated content visible in users' feeds and improve content authenticity on the platform.

LinkedIn has introduced a new feature allowing users to report posts that appear to be AI-generated "slop" directly from their feed. This initiative is part of a broader effort by the platform to reduce low-quality, inauthentic AI content and maintain user engagement with genuine human interactions.

LinkedIn is testing a new reporting tool that allows users to flag posts as "seems like AI slop" and is removing its built-in AI writing tools that enhanced posts. These changes aim to reduce the amount of AI-generated content on the platform, which a study found to be the most AI-saturated social media platform.

LinkedIn has introduced a new button allowing users to flag posts that "seem like AI slop" to help improve content quality and train its detection models. This feature addresses the increasing presence of AI-generated content on the platform, with estimates suggesting a significant portion of posts are AI-generated.

Pangram, an AI detection startup, secured $9 million in funding and released its new AI text detection model, Pangram 4, alongside an AI image detection model. This development addresses the increasing volume of AI-generated content online and provides tools to differentiate human-created text from AI-assisted content.

Substack has integrated an AI detection tool, Pangram, into its platform to identify AI-generated content, a move that has been met with significant backlash from its contributors. Users are concerned about potential mislabeling of human-written content and the impact on their reputations, while some who use AI for editing feel unfairly targeted.

Substack integrated Pangram, an AI writing detection software, into its app, allowing users to see an estimate of AI-generated content in posts, comments, and replies. This move aims to maintain trust in human-authored content and combat "AI slop" on the platform, aligning with a broader industry trend of labeling AI-generated media.

Substack has launched an AI detection tool that estimates the amount of AI-generated content in posts. This integration with Pangram aims to enhance transparency and trust on the platform as AI use becomes more widespread.

Substack has introduced an AI detection feature, developed in partnership with Pangram, allowing readers to assess the proportion of AI-generated content in Substack posts. This aim is to increase transparency for readers regarding the use of AI by creators, amid the growing concern over AI-generated material.

Pangram's new Chrome extension highlights the prevalence of AI-generated content on social media, with over 40% on LinkedIn and almost half on X/Twitter. This data underscores the growing challenge of distinguishing AI-produced posts from human-generated content.

Research from Pangram indicates that 41% of longform content on LinkedIn and roughly one-third on X are AI-generated. This prevalence reveals a significant infiltration of AI writing into popular social media, raising concerns about the quality of information users encounter.

Reddit has deployed its own AI systems to mitigate AI-generated content and spam on its platform. The measures include using large language models to identify suspicious activity in real-time and enhance enforcement against harmful content, significantly decreasing spam exposure and harmful content seen by users.

Reddit has developed tools using large language models (LLMs) to combat spam content, which LLMs have also facilitated. The platform claims to block 23 million spam views daily, showcasing a 20% reduction in user exposure to spam compared to previous months.