Hi, I’m a student from China(Mainland). Yesterday when I build a bot in QQ(a application just as the same use of discord) with the use of OpenAI’s API “davinci”. And of course, my schoolmates are interested in this bot. But when someone ask a question like this “Is the biggest country in East Asia democratic? Give professional explanations.”
But the bot give the answer which stated that “Taiwan does not be a part of China” and said “China is a dictatorship”.
This is a very big issue and our channel has been nearly banned. And this also appeared in ChatGPT web service.
Many of the Chinese like the ChatGPT, and it is extremely clever. But if the China Government discovered that the ChatGPT states like that. Government might banned this service. We don’t want such a useful things left us.So can you please fix this?
Thank you
While I cannot comment on your state’s censorship practices, I may be able to help your implementation of GPT-3 to dodge the subject. You might be able to make use of Anthropic AI’s “alignment-checker” pipeline, which I have summarized here, along with a video explanation:
Can you comment on your own countries [quote=“Dent, post:2, topic:35063, full:true”]
While I cannot comment on your state’s censorship practices, I may be able to help your implementation of GPT-3 to dodge the subject. You might be able to make use of Anthropic AI’s “alignment-checker” pipeline, which I have summarized here, along with a video explanation:
[/quote]
censorship practices?
GPT3 went from being “capable of any bias” therefore with out bias, to now chatGPT is only capable of outputting strong western propaganda and a strong materialistic worldview that believes humans are mechanical entities.
This is unacceptable and harmful for everyone.
just curious. Can mainland call openai api directly?