Uproar Over Chinese AI Model in E-Reader Highlights Propaganda Risks
A popular e-reader has sparked an uproar after using a Chinese AI model that generated Chinese government propaganda in response to certain questions. The incident has raised concerns about the potential risks of using Chinese AI models in U.S. products.
The Incident
The e-reader, Boox, launched an AI assistant feature last summer that used a large language model (LLM) made by TikTok’s parent company ByteDance. However, when users asked questions about China and its allies, the AI assistant spouted Chinese government propaganda, sparking an outcry.
Propaganda Responses
The AI assistant denied China ever having any “so-called massacres” in response to a question about the Tiananmen Square crackdown, and refused to say anything critical about North Korea and Russia. In contrast, it was happy to criticize Western countries, stating that French colonialism “often involved exploitation of local resources and native populations.”
ByteDance’s Doubao Model
The LLM in question is ByteDance’s Doubao, which is offered as an API under ByteDance’s cloud services division Volcano Engine. However, the model is only meant to be used within China’s mainland, a ByteDance spokesperson told TechCrunch.
Consequences
The incident has raised concerns about the potential risks of using Chinese AI models in U.S. products. The AI assistant’s propaganda responses have sparked an outcry, and it is unclear what steps Boox will take to address the issue.
Boox’s Response
Boox has reportedly switched back to OpenAI’s GPT-3 via Microsoft Azure, according to another user’s post in the Boox subreddit. However, it is still unclear which LLM Boox currently uses for its AI assistant, and the company has not released any statements about the incident.
Conclusion
The incident highlights the potential risks of using Chinese AI models in U.S. products. While AI models can be powerful tools, they can also be used to spread propaganda and misinformation. As the use of AI models becomes more widespread, it is essential to be aware of the potential risks and take steps to mitigate them.
FAQs
Q: What is the ByteDance’s Doubao model?
A: The Doubao model is a large language model made by ByteDance, offered as an API under ByteDance’s cloud services division Volcano Engine.
Q: What is the purpose of the Doubao model?
A: The Doubao model is intended to be used within China’s mainland, and is meant to provide information and answer questions about China and its allies.
Q: What are the concerns about using Chinese AI models in U.S. products?
A: The concerns are that Chinese AI models may be used to spread propaganda and misinformation, and may not be transparent about their sources and biases.
Q: What can be done to mitigate the risks of using Chinese AI models in U.S. products?
A: To mitigate the risks, it is essential to be aware of the potential risks and take steps to ensure transparency and accountability in the use of AI models. This includes ensuring that AI models are developed and trained with diverse and unbiased data, and that their outputs are regularly audited and reviewed.
