Archive — US — techcrunch
Reddit is using LLMs to solve a problem LLMs largely created
Reddit is utilizing large language models (LLMs) to combat the rise in spam, which has proliferated due to the same technology. The platform claims its new tools block 23 million spam views daily, reducing user exposure to spam by 20% in recent months. While LLMs enhance spam detection, experts stress the necessity of human moderation alongside AI efforts.
Key points
- LLMs have made it easier for bad actors to generate spam.
- Reddit has developed advanced tools using LLMs for spam detection and moderation.
- Reddit blocks 23 million spam views and identifies 25,000 new spam posts daily.
- The spam reduction tools show a 20% reduction in user exposure to spam from January to March.
- Other platforms are adopting similar AI strategies but emphasize human moderation as essential.
Why it matters
- The rise of AI-generated content is a significant trend in the tech landscape by 2025.
- Social media platforms are increasingly integrating AI to manage user-generated content and maintain community standards.
- The ironic situation of AI tools generating problems that they also help solve highlights ongoing discussions about ethical AI use.
Reddit's innovative use of LLMs to tackle an issue they helped create exemplifies the complex relationship between AI technology and its societal impacts, stressing the importance of combining AI capabilities with human oversight for effective moderation.