---
format: "aidr-story-markdown/v1"
id: "14e74690efc86c73294326d7139022f7004399500aba3e6033d756103c5e9b24"
canonical_url: "https://aidr.today/14e74690?lang=en"
title: "Meta Unveils TextSeal2 to Improve Detection of AI-Generated Text"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-02T16:00:47.000Z"
category: "Research"
topics: ["meta","llm","safety"]
source_urls: ["https://marketbrief.now/ai/meta-unveils-textseal2-to-improve-detection-of-ai-generated-text-8c78c53f","https://huggingnews.com/ai/meta-unveils-textseal2-to-improve-detection-of-ai-generated-text-8c78c53f"]
summary: "Meta unveiled TextSeal2 to improve detection of AI-generated text. The TextSeal2 system uses \"copula mixing\" to embed a hidden signal in every token. In benchmarks using a Qwen3.5-9B model, TextSeal2 identified 89% of 200-token responses at a 0.1% false-positive rate."
---

# Meta Unveils TextSeal2 to Improve Detection of AI\-Generated Text

> [Open the canonical story](<https://aidr.today/14e74690?lang=en>)

**Published:** 2026-11-02T16:00:47.000Z
**Category:** Research
**Topics:** meta, llm, safety

## Summary

Meta unveiled TextSeal2 to improve detection of AI\-generated text\. The TextSeal2 system uses "copula mixing" to embed a hidden signal in every token\. In benchmarks using a Qwen3\.5\-9B model, TextSeal2 identified 89% of 200\-token responses at a 0\.1% false\-positive rate\.

## Sources

- [Story source](<https://marketbrief.now/ai/meta-unveils-textseal2-to-improve-detection-of-ai-generated-text-8c78c53f>)
- [Story source](<https://huggingnews.com/ai/meta-unveils-textseal2-to-improve-detection-of-ai-generated-text-8c78c53f>)