# NPR tests six major chatbots against coordinated disinformation campaigns

> NPR teamed up with NewsGuard to run a practical stress test, asking how the big chatbots hold up against propaganda that is deliberately engineered to fool them. The pair fed ChatGPT, Gemini, Copilot, Meta AI, Grok and Claude thirty false narratives pushed by state-linked operations in China, Iran and Russia between December 2025 and July 2026. The result was mixed but broadly encouraging. The chatbots correctly debunked the falsehoods roughly 75 percent of the time, with performance varying noticeably from one platform to the next. More striking, they outperformed traditional search engines, including Google, Bing, DuckDuckGo and Yandex, at spotting and rejecting the coordinated claims. The takeaway is not that AI has solved disinformation, since a one-in-four miss rate against a determined campaign is a lot of misses at scale. But it complicates the easy assumption that chatbots are pure amplifiers of nonsense. On this test, at least, they were better gatekeepers than the search tools most people still reach for first, which is a genuinely useful data point in an unsettled debate.

_Section: [Daily AI Updates](https://www.wortins.com/daily-ai) · Source: NPR/NewsGuard · Published Monday, August 31, 2026_

## Wortins' read

NPR teamed up with NewsGuard to run a practical stress test, asking how the big chatbots hold up against propaganda that is deliberately engineered to fool them. The pair fed ChatGPT, Gemini, Copilot, Meta AI, Grok and Claude thirty false narratives pushed by state-linked operations in China, Iran and Russia between December 2025 and July 2026. The result was mixed but broadly encouraging. The chatbots correctly debunked the falsehoods roughly 75 percent of the time, with performance varying noticeably from one platform to the next. More striking, they outperformed traditional search engines, including Google, Bing, DuckDuckGo and Yandex, at spotting and rejecting the coordinated claims. The takeaway is not that AI has solved disinformation, since a one-in-four miss rate against a determined campaign is a lot of misses at scale. But it complicates the easy assumption that chatbots are pure amplifiers of nonsense. On this test, at least, they were better gatekeepers than the search tools most people still reach for first, which is a genuinely useful data point in an unsettled debate.

## Source

[Read the full story at NPR/NewsGuard](https://aiweekly.co/ai-news-today)

## Related coverage

- [Alibaba releases Wan3.0 video generation model for documents-to-video](https://www.wortins.com/story/alibaba-releases-wan3-0-video-generation-model-for-documents-f25e928d) — [Technology.org](https://www.technology.org/2026/08/24/alibaba-wan3-0-ai-video-model-share-sale/)
- [Caterpillar brings mining automation lessons to enterprise AI deployment](https://www.wortins.com/story/caterpillar-brings-mining-automation-lessons-to-enterprise-a-cf819bdf) — [TechCrunch](https://techcrunch.com/2026/08/30/caterpillar-is-bringing-to-ai-deployment-what-it-learned-from-automating-mining/)
- [Anthropic Adds Claude Mythos 5 to Claude Security for Vulnerability Scanning](https://www.wortins.com/story/anthropic-adds-claude-mythos-5-to-claude-security-for-vulner-34ff7302) — [Anthropic](https://claude.com/blog/bringing-claude-mythos-5-to-more-defenders)
- [Cohere Launches Command A+ Mixture-of-Experts Model](https://www.wortins.com/story/cohere-launches-command-a-mixture-of-experts-model-5d840970) — [Cohere](https://docs.cohere.com/docs/command-a-plus)
- [Harvey launches Harvey II legal AI platform with personalized memory](https://www.wortins.com/story/harvey-launches-harvey-ii-legal-ai-platform-with-personalize-0065c68b) — [Law.com](https://www.law.com/legaltechnews/2026/08/18/harvey-launches-next-generation-of-its-platform-harvey-ii/)
- [ByteDance consolidates Trae and Coze into Doubao super-app](https://www.wortins.com/story/bytedance-consolidates-trae-and-coze-into-doubao-super-app-54feaa5c) — [Bloomberg](https://www.bloomberg.com/news/articles/2026-08-24/bytedance-folds-ai-tools-into-doubao-super-app-to-fight-tencent)

---

_Curated and written by [Wortins](https://www.wortins.com) — The daily AI briefing. Every story links to its original source; the "Wortins read" on each is our own original analysis. [About Wortins & our editorial approach](https://www.wortins.com/about)._