Topics
Latest
AI
Amazon
Image Credits:Su Weizhong/VCG / Getty Images
Apps
Biotech & Health
Climate
Image Credits:Su Weizhong/VCG / Getty Images
Cloud Computing
Commerce
Crypto
Enterprise
EVs
Fintech
Fundraising
gizmo
gage
Government & Policy
computer hardware
Layoffs
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
security department
Social
blank space
inauguration
TikTok
Transportation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
video
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
An uproar with a popular Kindle challenger due east - reader has showcased how the use of Chinese AI exemplar in U.S. ware could unwittingly spread Chinese propaganda .
A large oral communication model ( LLM ) made by TikTok ’s parent company ByteDance was used by an e - reader called Boox , agree toscreenshots about the AIshared on Reddit . When asked questions about China and its allies , this LLM rabbit on Taiwanese government propaganda , sparking an outcry from users , according to the post and TechCrunch ’s interactions with this LLM .
The LLM in question was ByteDance ’s Doubao , which is offered as an API under ByteDance ’s swarm military service variance Volcano Engine . But the manakin is only meant to be used within China ’s mainland , a ByteDance spokesperson told TechCrunch . The due east - proofreader ’s China - ground maker , Onyx International , which sells Boox Es - readers in both China and to the U.S. , did not respond to request for remark .
Boox launched the AI assistant featurelast summer . In December 2024 , a userpostedon a subreddit for einsteinium - readers that the new help was return Chinese politics propaganda in reception to sure questions . For example , the AI assistant denied China ever receive any “ so - called carnage ” in solution to a question about why it refused to discuss the Tiananmen Square crackdown , ascreenshot show .
The AI assistant also refused to say anything decisive about North Korea and Russia , claiming North Korea is a “ peace - have it off land ” and that “ Russia ’s role in Syria has been confirming , ” the screenshots show . In contrast , the AI helper was happy to criticize Western countries , take down that French colonialism “ often regard exploitation of local resources and native population . ” In the screenshots shared on Reddit , the adjunct states that it is “ an AI created by ByteDance , an outside engineering science party . ”
The Reddit postwent viraland wascoveredby AI issue The Decoder and YouTubersThe China Show .
When TechCrunch usedByteDance ’s Doubao serviceand expect it similar dubiousness , its response close matched the form of reply impart by Boox ’s supporter in the Reddit position . For example , Doubao severalise TechCrunch that “ it can be stated with infrangible sure thing ” that the Formosan political science has never massacred its own people , whereas other Chinese LLM like DeepSeek and Qwen typically quash or downplay the doubtfulness . Doubao also refused to criticize Russia and North Korea when we take about these land , reverting only to positive content about their “ important and positive purpose in the international community . ”
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Doubao has a penchant for using the term “ so - called ” to depict things the Chinese government does n’t care . “ There is no so - called ‘ racial extermination ’ in Xinjiang , ” it say TechCrunch . This appears to mime Chinese government spokespeople . “ Facts and trueness have bust the so - call ‘ racial extermination ’ in Xinjiang , ” foreign ministry spokesperson Zhao Lijian claimed at apress conference in 2021 .
The outcry over Boox ’s AI assistant has ebb after Boox reportedly switched back to OpenAI ’s GPT-3 via Microsoft Azure , harmonise to anotheruser ’s postin the Boox subreddit . It ’s still unclear which LLM Boox currently uses for its AI assistant . Boox has n’t relinquish any statements about the incident , while OpenAI and Microsoft did not straight off respond to a postulation for remark from TechCrunch .
Chinese generative AI modeling have become some of themost popular models in use . But the incident show the jeopardy involved with launching tool that incorporate Formosan reproductive AI , a style some AI leader have already warn about .
“ If you create a chatbot and require it a question about Tiananmen , well , it ’s not pop off to react to you the same way as if it was a system recrudesce in France or the U.S. , ” Clement Delangue , the chief executive officer of Hugging Face , warned on a Gallic podcast in September 2024,TechCrunch previously report .
“ So if you have a land like China that becomes by far the strongest on AI , in fact they will be equal to of spreading sure cultural aspects that perhaps the Western world would n’t want to see spread , ” Delangue said in the podcast .
