Last year, Scientists at Duke University published PULSE, an algorithm that clarifies blurry photos. But a blurry photograph of former president Barack Obama was altered to produce a clear white face.
This is seen as a classic case of algorithmic bias. Algorithmic bias is the influence of non-neutral position on users in the process of information production, distribution and verification, which leads to the spread of one-sided and untrue information concepts. Algorithmic bias can happen to everyone in the real world. Wallach (2016) affirmed that as long as the bias existed in the society, more attention is given by machine learning system on this specific bias. In the era of capitalised platform, the ideology and value transmitted are mastered and controlled by the ruling class and head of social platform. The discourse of racism and patriarchal are dominant in the previous online communities and drive platforms, which shows the biases on feminism, anti-racism or other social topic, into public controversy. According to Google (2017), the biases manipulated by platforms are based on Interaction Bias, Potential Bias and Selection Bias. The process of machine learning still has a long way to go.
Furthermore, I am going to critically analyse the content of Chinese social media named Weibo within a feminism perspective by narrowing down the content on platform to one specific section.
Hot-spot Topic List：
The 11th topic (A company executive in Shanxi was suspended for domestic violence to his wife) was the sign of feministic revolution on Chinese social media and reversing stereotypical policy of Chinese companies. As Weibo is the first feminists-dominated platform in China, this topic has been pushed to the hot-spot list immediately since the event was exposed. However, the 12th topic (Leung advises women to choose a spouse who is not necessarily stronger than they are) is quite controversial among Weibo users. The suggestion to women would enhance the stereotype that women always desire to find a boyfriend or partner who is more competent. It opposed the common value among female citizens of shaping a better self but not dependence on men. Most feministic discussers claims that the topic was raised because women are becoming stronger in the society and men, on the contrary, are asking women to accept partners weaker than themselves because their pride doesn’t allow them to be the weak one. In fact, men are the ones who need to become stronger.
The topic of gender equality changes from a marginal issue to a mainstream issue on Weibo in recent 10 years. Since 2009, Sina Weibo has been the main internet platform for public events to take place, and its development has been a continuous process of regulation and commercial logic infiltrating the social space created by technological innovation.
Due to the trend of commercialised feminism and increasing amount of female users on Weibo，the algorithm on the platform tend to matches the female audiences. Meanwhile, the gender discourse and justices published by feministic users influence and restructure machine learning system of Weibo.