An algorithmic bias in social networking sites has more potential for creating chaos in the world.

Social networking sites are used for many purposes, apart from posting about one’s life. In recent times, social networking sites have been admittedly utilized by youngsters to be aware of the incidents happening around them and the world. In fact, the news agencies and sites are utilizing the power of social media to reach the audience, especially youngsters, so that they can be aware of the things that might potentially affect them.

With its interface, that connects people, helps media and news organizations to connect with people, and enables many brands to advertise their products, one of the flaws associated with social networking sites that have been discounted over the years is the algorithmic biases. Enough evidence has been retained about almost all the social networking sites’ algorithmic biases that it becomes imperative for the world to acknowledge that.

Undoubtedly, systemic biases are implemented in the algorithm of social networking sites unconsciously. But it is these systemic biases that the world is fighting, even before the social networking sites’ commemoration. Many organizations, societies, and social welfare groups have raised brows towards why systemic biases exist, but no definite answer has been delivered about this despite this.

Owing to their algorithm’s systemic biases, these social media sites have been subjected to many lawsuits. In a very famous incident, the Facebook was required to pay US$5 billion to the Illinois Facebook users for infringing their privacy by collecting data using facial recognition technique between 2015 to 2020, without consulting or informing them. The lawsuit was filed against infringing the Biometric Information Privacy Act.

However, despite the lawsuits, social networking sites are unable to mitigate the events of systemic biases.

Racial bias algorithm of Twitter 

Twitter has been, perhaps, the only social media networking site, which permits its users to directly interact with their favorite celebrity, sportsperson, and politicians. Also, twitter is the only social network site, which provides the platform for its users to get notified about the world’s events.

On September 20, the twitter users were shocked to witness the algorithmic bias flouting in the social networking site, which cropped the images of dark skin toned users to display white people’s face. As the company is re-evaluating the cause of this algorithmic bias, reports suggest that this bias has been apparent since 2017, when Twitter scrapped out its facial detection bias and replaced it with more salient detection features allowed the auto-cropping of the images. In its blog post in 2017, a new promulgated algorithm aimed to crop the users’ image, focusing on “Salient” image features. It will also allow the users to utilize the data for training neural networks and algorithms in predicting the image of the user and cropping the most exciting region of the image, as per the users’ choice. This was done to rectify the facial recognition bias, which was earlier integrated with twitter.

However, with the pre-existing issue of Twitters’ algorithmic bias getting observed again, experts believe that despite installing a new saliency algorithm, the ingrained cause of such biases has not been solved.

The algorithmic bias researcher, Vinay Prabhu, has formulated a methodology for assessing the cropping bias algorithm of Twitter. However, Zehan Wang, a twitter engineer, said in his tweet, “We’ll look into this. The algorithm does not do face detection at all (it replaced a previous algorithm which did). We conducted some bias studies before release back in 2017. At the time, we found that there was no significant bias between ethnicities (or genders).”

Earlier this year, the Twitters’ Chief in India was arrested for showing selectiveness in removing hate-speech. A Wall-Street Journal investigation revealed callousness while deleting posts and accounts of Hindu Nationalists who made remarks against the minorities.

But this is not the first time that Twitter is under the radar for its inherent systemic bias.

Facebooks’ gender and race-biased algorithm

In 2019, a study was conducted by the researchers of Northwestern University regarding specific tweaks in the ad-algorithms of Facebook. It was found that those tweaks had a significant impact on the jobs pertaining to minorities and women. For example, it was observed that for the job postings of Janitors and taxi drivers, the higher number of people belonging from the minority community were notified about the job, at the same time, for the job postings of nurses and teachers, women were mostly shown the advertisement.

The Social Networking site was already sued by the US Department of Housing and Urban Development for purposely targeting people based on race, gender, and religion regarding the advertisements.

In early September this year, a study was conducted by a UK-based counter-extremist organization, Institute for Strategic Dialogue, which found that facebook’s algorithm was actively promoting Holocaust Denial Content. The investigation revealed that facebooks’ algorithm promotes 36 holocaust denial pages, with more than three lakh followers. A similar investigation was also carried out on other social media platforms, that revealed 2,300 pieces on Reddit, 19,000 on twitter, and 9,500 on youtube promoting holocaust denial.

Since 2016, the involvement of Facebook, has been observed for tweaking results in some of the major events. Be it during the UK referendum for Brexit or the USA elections 2020; the social networking site has been observed with skepticism.

Pinterest’s Gender Bias Lawsuit

This year, Pinterest, another social networking site, was in the scrutiny of federal authorities regarding its gender bias, when the company’s former Chief Operating Officer Francoise Brougher, filed a lawsuit against the company citing Gender bias, and that she was offered less favorable pay structure. Even though this has been nothing technologically alarming, but this signifies the systemic bias that has been ingrained in a high-ranked social networking site like Pinterest.

Instagrams’ Address ofAlgorithm

Owing to an increase in momentum of the ongoing #BlackLivesMatter protest, Instagram, in a blog post, shared that it will scrutinize its algorithm so that no racial content can be generated, both with regard to “Shadowbanning” and structural biases in its system. Shadowbanning refers to “filtering people without transparency, and limiting their reach as a result,” as described by Instagram.

Earlier, Facebook has also said to focus on four issues- Harassment, Verification, Distribution, and Algorithmic bias, to mitigate the occurrence of systemic biases.

Conclusion

As social media is the first source of information these days, an inherent systemic bias carries a higher potential for misinformation, falsification of facts, fake news, racial discrimination, and gender biases. It will have a negative impact on society. Hence the social networking sites must access their algorithms so that such incidents do not happen in the future. The world is already grappling with vanquishing the societal biases and stereotypes. With social media algorithms, entertaining the cognitive and systemic biases, the struggle and the desired outcome will only be delayed.