A few days ago, a whistleblower from Facebook named Frances Haugen made public a report that exposed startling information about how the company has responded to hate speech. Haugen revealed that, despite its efforts, Facebook only managers to remove about 3-5% of the hate speech that occurs on its platform, according to its own estimates. While this statistic is troubling in itself, it is made worse when you consider what kind of impact such language can have on the people that use the site.
Profit vs. protection
When publicizing the report, Haugen commented “I don’t trust that they’re willing to actually invest what needs to be invested to keep Facebook from being dangerous”. In other words, Facebook has been accused of putting profit over safety on its platform. While investing in stopping hate speech is good for society, it is not necessarily good for Facebook’s bottom line, as hateful content can drive more people to spend more time on the site.
However, this is starting to change. A group of advertisers, including the North Face and Patagonia, are demanding that Facebook ramp up their efforts to reduce the presence of hate speech on the site. As part of this campaign, titled “Stop Hate for Profit”, the advertisers are refusing to collaborate with Facebook until it takes actionable steps to improve the issue. Efforts such as these could make allowing hate speech costly, potentially tipping the scales and hopefully encouraging Facebook to make a change.
Where do we go from here
While many argue that legal regulation is the best way to reduce hate speech online, this is not as easy as it seems. For one, invoking regulations in this area is difficult without infringing on the right to free speech. Furthermore, catching such speech on a large scale would require intensive investment in text analytics and employees to examine flagged posts and discern what constitutes free speech. As a result, it may take a long time before we begin to see progress towards eliminating hate speech online.
https://www.theguardian.com/technology/2020/jun/22/facebook-hate-speech-advertisers-north-face
Interesting read! The problem for Facebook is that removing hate speech needs to be done by humans, as it is nearly impossible for AI to detect what constitutes hate speech. Language is all about context.
It has been clear in the past that Facebook’s efforts towards removing hate speech has been quite lackluster. I remember this article from Reuters that described how Facebook had become the platform through which prominent figures in Myanmar share their hate of the Rohingya, a minority group in the country. Meanwhile, Facebook only had two Burmese speaking employees that reviewed content from Burmese speaking countries. Only after the U.S. Congress inquired with Zuckerberg did Facebook start hiring more Burmese speaking content reviewers. That is quite telling of how Facebook operates.
https://www.reuters.com/investigates/special-report/myanmar-facebook-hate/