Findings of the first Facebook-funded regional report into hate speech regulation show that, despite automated moderation and improved content standards, the platform is still allowing vilification and discrimination to proliferate on public pages, with inadequate mechanisms to address it.
The report, Regulating Hate Speech in the Asia Pacific, analysed a cross-section of public pages administered by LGBTQI+ groups in the Philippines, India, Indonesia, Myanmar and Australia, and found most felt they had little or no support from the platform when they were targeted by hate speech.
In India, Indonesia and the Philippines in particular, LGBTQI+ groups are exposed to an unacceptable level of discriminatory, hateful and threatening posts, which escape Facebook’s algorithmic and human filters.
Administrators of these pages report receiving inconsistent levels of help from Facebook when they report hate speech and seek to have posts taken down. They often receive automated rejections and some posts have even reappeared after a few days.
The report authors from the universities of Sydney and Queensland are calling on Facebook to take a more pro-active role in moderating and combatting hate speech in consultation and cooperation with regional ethnic and minority community groups across the region.
“Hate speech evolves so fast that algorithms struggle to keep up. Facebook needs to consult more regularly and widely with the groups that are most impacted by hate speech including LGBTQI+ and women’s groups, Muslims and ethnic minorities,” said report co-author Associate Professor Fiona Martin, a leading expert in social media trends from the Department of Media and Communications at the University of Sydney. “An annual Asian region hate speech mitigation event with all stakeholders would be ideal.”
“We think it’s important that anyone who runs a public Facebook page should get mandatory training in how to identify and report hate speech,” said Associate Professor Martin. “Our research suggests that page admins are not always aware of Facebook’s advice pages, and only a few received training in responding to hate posts from civil society groups.”
The report also encouraged Facebook to provide more transparent, better supported reporting mechanisms.
It suggests that those on the frontline of managing public pages for such groups in Asian countries want Facebook to do more to help them mitigate hate speech, where they are suffering the damaging effects of harmful content on a regular basis.
Lead author Dr Aim Sinpeng, senior lecturer in comparative politics at the University of Sydney, said: "We were surprised to find that Myanmar LGBTQI+ groups on Facebook experienced very little hate speech given that Facebook has been condemned for contributing to genocide against the Rohingyas.
This may be because Facebook has worked with the UN and other groups to specifically tackle hate speech in that country. It has also expanded its definition of hate speech and is working towards moderating in more languages which is having a positive impact on harmful content.
"We were also surprised to find that there is no correlation between religious beliefs and the level of hate speech experienced on Facebook by LGBTQI+ groups in a country,” said Dr Sinpeng. “Instead, our findings show that in cases where the issue of LGBTQI+ rights is politicised, we see more hate speech on social media."
Dr Sinpeng said the research team is concerned that the COVID-19 pandemic has made Facebook more reliant on automation when it comes to hate speech moderation.
This report was funded as part of the Facebook Content Policy Research on Social Media Awards to examine Facebook’s hate speech regulation challenges in the Asia Pacific.
It maps hate speech law in five case study countries in the Asia Pacific region - India, Myanmar, Indonesia, the Philippines and Australia - to understand how this problem is framed nationally, and what regulatory gaps exist that might enable hate speech to proliferate on the platform.
Finally, given the level of discrimination and vilification experienced by lesbian, gay, bisexual, transgender, intersex and queer identifying people across Asia, the study analyses data from the public Facebook pages of major LGBTQI+ groups in the case study countries to examine the incidence of hate speech that escaped Facebook’s automated filters.
It also presents interviews with page administrators of these groups that reveal their conceptions and management of hate speech, including their experience of reporting hate posts to Facebook.
Along with expert online community management input from the Australian Community Managers network, these interviews provide a framework for understanding the ‘regulatory literacy’ of those who are at the frontline of Facebook’s efforts to minimise hate on its platform.
This is the first team of Australian social scientists to be awarded Facebook funding for social media research, and the first integrated, comparative study of hate speech on Facebook in many countries across the region.
Declaration: The research team received a grant from Facebook to conduct this study, which was in the form of an “unrestricted research gift.” The terms of this gift specify that Facebook will not have any influence in the independent conduct of any studies or research, or in the dissemination of findings. Top image: Unsplash