DeepSeek launched a free, open-source large language model in late December, claiming it was developed in just two months at a cost of under $6 million.
I mean that’s the kind of answer DeepSeek gives you if you ask it about Uyghurs. “Some say it’s a genocide but they don’t so guess we’ll never know ¯_(ツ)_/¯”, it acts as if there’s a complete 50/50 split on the issue which is not the case.
So you expect that an AI provides a morally framed view on current events that meet your morally framed point of view?
The answer provides a concise overview on the topic. It contains a legal definition and different positions on that matter. It does at not point imply. It’s not the job of AI (or news) to form an opinion, but to provide facts to allow consumers to form their own opinion. The issues isn’t AI in this case. It’s the inability of consumers to form opinions and their expec that others can provide a right or wrong opinion they can assimilation.
I agree and that’s sad but that’s also how I’ve seen people use AI, as a search engine, as Wikipedia, as a news anchor. And in any of these three situations I feel these kind of “both sides” strictly surface facts answers do more harm than good. Maybe ChatGPT is more subtle but it breaks my heart seeing people running to DeepSeek when the vision of the world it explains to you is so obviously excised from so many realities. Some people need some morals and actual “human” answers hammered into them because they lack the empathy to do so themselves unfortunately.
I mean that’s the kind of answer DeepSeek gives you if you ask it about Uyghurs. “Some say it’s a genocide but they don’t so guess we’ll never know ¯_(ツ)_/¯”, it acts as if there’s a complete 50/50 split on the issue which is not the case.
So you expect that an AI provides a morally framed view on current events that meet your morally framed point of view?
The answer provides a concise overview on the topic. It contains a legal definition and different positions on that matter. It does at not point imply. It’s not the job of AI (or news) to form an opinion, but to provide facts to allow consumers to form their own opinion. The issues isn’t AI in this case. It’s the inability of consumers to form opinions and their expec that others can provide a right or wrong opinion they can assimilation.
I agree and that’s sad but that’s also how I’ve seen people use AI, as a search engine, as Wikipedia, as a news anchor. And in any of these three situations I feel these kind of “both sides” strictly surface facts answers do more harm than good. Maybe ChatGPT is more subtle but it breaks my heart seeing people running to DeepSeek when the vision of the world it explains to you is so obviously excised from so many realities. Some people need some morals and actual “human” answers hammered into them because they lack the empathy to do so themselves unfortunately.
\ here you dropped an arm
¯\_(ツ)_/¯
If you verbose, you can see all the reasoning behind the answers. With Taiwan, it’s hard coded in without /thinking