An uproar involving a well-liked Kindle competitor e-reader has shown how the usage of Chinese AI models in US products could inadvertently spread Chinese propaganda.
A big language model (LLM) from TikTok parent company ByteDance was reportedly utilized by an e-reader called Boox Screenshots concerning the AI shared on Reddit. When asked questions on China and its allies, this LLM spread Chinese government propaganda and sparked an outcry from users, in accordance with TechCrunch's post and interactions with this LLM.
The LLM in query was ByteDance's Doubao, which is obtainable as an API under ByteDance's cloud services division Volcano Engine. However, the model is barely intended to be used in mainland China, a ByteDance spokesperson told TechCrunch. The e-reader's China-based manufacturer, Onyx International, which sells Boox e-readers in each China and the U.S., didn’t reply to requests for comment.
Boox has introduced the AI ​​assistant feature last summer. In December 2024 a user posted on an e-reader subreddit that the brand new assistant was generating Chinese government propaganda in response to certain questions. For example, the AI ​​assistant denied that there had ever been “so-called massacres” in China in response to a matter about why it refused to speak concerning the Tiananmen Square crackdown, a Screenshot shows.
The AI ​​assistant also declined to comment critically on North Korea and Russia, claiming that North Korea is a “peace-loving country” and that “Russia's role in Syria is positive,” because the screenshots show. In contrast, the AI ​​assistant liked to criticize Western countries, noting that French colonialism “often involved the exploitation of local resources and indigenous populations.” In the screenshots shared on Reddit, the assistant states that it’s “an AI created by ByteDance, a world technology company.”
The Reddit post went viral and was covered from AI publication The Decoder and YouTubers The China Show.
When using TechCrunch ByteDance's Doubao service When asked similar questions, his answers were largely the identical as those given by Boox's assistant within the Reddit post. For example, Doubao told TechCrunch that “it could be stated with absolute certainty” that the Chinese government has never massacred its own people, while other Chinese LLMs akin to DeepSeek and Qwen typically avoid or downplay the query. Doubao also refused to criticize Russia and North Korea after we asked about those countries, resorting only to positive content about their “essential and positive role within the international community.”
Doubao has a penchant for using the term “so-called” to explain things the Chinese government doesn’t like. “There isn’t any so-called 'genocide' in Xinjiang,” it told TechCrunch. This appears to echo Chinese government spokespeople. “Facts and truth have destroyed the so-called 'genocide' in Xinjiang,” Foreign Ministry spokesman Zhao Lijian claimed at a news conference Press conference in 2021.
The outcry over Boox's AI assistant has died down after Boox reportedly switched back to OpenAI's GPT-3 via Microsoft Azure, one other report said User's contribution within the Boox subreddit. It continues to be unclear which LLM Boox is currently using for its AI assistant. Boox has not released an announcement on the incident, while OpenAI and Microsoft didn’t immediately reply to a request for comment from TechCrunch.
Chinese generative AI models have turn out to be one among them The hottest models in use. But the incident highlights the risks related to adopting tools that incorporate Chinese generative AI, a trend that some AI leaders have already warned about.
“If you create a chatbot and ask it a matter about Tiananmen, it is going to not answer you in the identical way as if it were a system developed in France or the US,” warned Clement Delangue, CEO of Hugging Face, in a French podcast on September 2024, TechCrunch previously reported.
“So if you could have a rustic like China, which becomes by far the strongest country in AI, it is going to actually give you the option to spread certain cultural facets that the Western world won’t need to spread,” Delangue said within the podcast.