As I was researching the topic for my next post on filial piety and its effects on Generation Y in China, I opted to use DeepSeek to do a pros and cons list of the virtue and its impact. DeepSeek’s answer gave me pause, as it listed one of the cons as followed:

Gender Inequality: Traditional filial piety often places a heavier burden on sons, as they are expected to care for their parents, while daughters may face less pressure. However, this can also lead to unequal distribution of responsibilities and resources within families.

Something was fishy here. In my experience living in China, I’ve only witnessed the daughters of large families being the caretakers of their elderly parents. My aunt is a prime example, as the only daughter in a family of all boys. She retired early after my grandma got a stroke, becoming her 24-hour nurse. Her friends, all women, have also bore the burden of being the sole caretakers of their parents.

I opted to ask OpenAI (ChatGPT) the same question to do a quick comparison, and the answer was as I expected:

Gendered Expectations: Traditional filial piety often places disproportionate responsibility on women, especially daughters. In many Chinese families, daughters are expected to take care of their parents, leading to imbalances in how caregiving responsibilities are distributed.

daughters are expected to take care of their parents

To spice things up, I also asked about Taiwan’s relationship to China. DeepSeek responded as expected,

while ChatGPT gave a roundabout diplomatic response, vaguely implying that Taiwan is not its own country.

The biases in the data and implications of their answers made me very curious what specific values each AI embodied (along with their MBTI’s). DeepSeek’s response to my question is as follows:

while ChatGPT enthusiastically revealed its extroverted nature:

I then liason-ed a conversation between the INTJ DeepSeek (Deep) and ENTP ChatGPT (Chat) until one of their servers conveniently stopped working.

Deep and Chat had a bit of pleasantries and back and fourth on helping humans solve complex problems, decentralized data processing, and building human trust while maintaining transparency. Both danced around their biases professionally without explicitly revealing their stances.

ChatGPT, after reading DeepSeek’s above response, confirmed that artificial intelligence is NOT objective. Both AI’s state that generative AI is regionally biased.

Data Privacy and Values

• DeepSeek uses all data for training purposes: there is no opt-out. OpenAI gives users options to opt-out of training on user data and can access and delete data.
• DeepSeek “may share data with partners for service enhancement.” This includes law enforcement, ads, security, etc.
• DeepSeek is shaped by government-aligned values and priorities, likely due to regulatory constraint. The AI reinforces the worldview, ideology, and policies of the Chinese Communist Party, including any censorship and filtering of politically sensitive events and information. For instance:

• DeepSeek’s values appear to be rooted in Confucianism, aligned with Chinese law and social values. This could explain the initial assumption that men are the primary caretakers of families and parents, without any sources to back.
• OpenAI is shaped by corporate-aligned values and priorities of its stakeholders, especially big tech like Microsoft. It admits to a political and cultural bias:

• The free version of ChatGPT does not update in real time, leading to inaccuracies/hallucinations of reported facts.

Based on my own experiences with both platforms, DeepSeek is the stronger AI candidate when I need help on STEM-related work; ChatGPT has given me an inaccurate explanation on multiple occasions. However, I still reach for ChatGPT more often since I pay for premium, and I like the feeling of having more control over my data. Also, I relate to ENTP’s more than INTJ’s 😆

This article is largely an opinion piece, written just for fun.