E-Paper
Meta global policy management head interview: Spending over $5 billion this year on user safety
“The truth is we’re on track to spend more than $5 billion this year alone on safety and security and have over 40,000 people to do one job: keep people safe on Facebook,” Meta’s head of global policy management Monika Bickert told HT in an interview.
Meta, previously known as Facebook, continues to be mired in controversy surrounding its allegedly partisan policies when it comes to moderating speech, combating inappropriate online behaviour, and addressing fake news. The company’s head of global policy management Monika Bickert, in written responses to queries from HT claimed that the company ensures users’ safety on its platforms and is on track to spend over $5 billion dollars this year on safety and security. She also argued that the premise of the allegations made by former company employees and whistleblowers, such as Frances Haugen and Sophie Zhang, are false as the firm doesn’t compromise user well-being for profits. Edited excerpts
Facebook recently rebranded itself as Meta, what all are the changes that will follow the transformation?
We’ve always believed that the ultimate promise of technology is to improve people’s lives, and bring people closer together. Now, we see the opportunity on the horizon to bring people together in ways never before possible. The Metaverse is the next frontier of virtual social connection. This is where we can make the greatest contribution -- in helping to develop the next generation of online social experiences. We’re already building the foundational technologies needed to help deliver this AR/VR future -- our work on communities, creators, commerce, and the next computing platform are the building blocks for a completely new generation of experiences. We are at the beginning of something very important in the history of communication, and we plan to be at the forefront, helping to lead the way. The last decade saw a shift from desktop web to mobile internet and the next decade will be about shifting into the metaverse. Because the metaverse will be inherently social and will connect people in ways never before thought possible, it is the fulfilment of our mission.
How does Meta respond to the allegations made by several whistleblowers, such as Frances Haugen and Sophie Zhang, that its policies are detrimental to its users?
At the heart of these allegations is a premise which is false. Yes, we’re a business and we make profit, but the idea that we do so at the expense of people’s safety or well-being misunderstands where our own commercial interests lie. The truth is we’re on track to spend more than $5 billion this year alone on safety and security and have over 40,000 people to do one job: keep people safe on Facebook.
Has Facebook affected any changes, especially in India, since the allegations and documents became public?
Billions of people use our services because we build the best tools to stay connected to the people they care about, to find communities that matter to them, and to grow their small business. And the reason we’ve been able to succeed for almost two decades is because we keep evolving and building.
Our vision has always been to foster an ecosystem of apps that create a richer and more interesting experience for people. At the same time, we have developed rules to make our apps work better for everyone and to protect the safety and privacy of people who use them.
People, communities, and businesses in India continue to join our platforms every day as they see the value -- the positive ways in which we empower and touch their lives. We’ve made changes to our policies over time to deal with issues as they arose. We stand by those changes and will continue to evaluate our policies to address any new issues that arise. We’re committed to continuing our industry leading work around user safety.
How exactly do you identify hate content and inauthentic accounts in a country as diverse as India, where several languages and dialects are used on the social media platform?
Our aim has been to remove abuse from our platform. Over the years we have invested significantly in technology to find hate speech and tackle misinformation on our platform. Since 2016, we have invested more than $13 billion on our teams and technology that focuses on safety and security and this year we are on track to spend more than $5 billion in this area. Today, we have tripled the number of people working on safety and security issues to more than 40,000, including more than 15,000 dedicated content reviewers who review content in more than 70 languages. This team is based in over 20 sites globally, including India and covers every major time zone. For India, we have people reviewing potentially problematic content in 20 Indian languages including Assamese and we have hate-speech classifiers in four Indian languages -- Hindi, Bengali, Tamil and Urdu.
What is the role of the India policy team? How often does the company receive feedback from them?
The India public policy team is responsible for engaging with policymakers, civil society and other policy stakeholders on policy development initiatives around a host of topics such as user safety, data protection and privacy and internet governance to shape the future of technology and represent the company’s position in these multi-stakeholder processes.
Does the India team have any role in content moderation? Several allegations have been directed at a former public policy head of the Indian team for failing to moderate content because of political considerations?
We make decisions on content if they violate our policies, not because of someone’s political position, party affiliation, or political point of view. Our Content Policy team that writes the policies on what is and isn’t allowed on our platforms, operates independently from our Public Policy team, who engage with policymakers. When updating our policies or when looking for input in escalating difficult decisions, Content Policy relies on input from many organizations across the company (including Operations, Engineering, Legal, Human Rights, Civil Rights, Safety, Comms and Public Policy).
In these instances, Public Policy is just one of many groups consulted. And, while the perspective of Public Policy is key to understanding local context, no single team’s opinion has more influence than the other. We also have a separate team who enforces these policies.
Facebook has often asserted that it remains free of political interference, but have there been instances wherein it has had to alter its stance for fear of political retaliation?
Our Community Standards apply to everyone, all around the world, and to all types of content. We enforce these policies globally without regard to anyone’s political position, party affiliation or religious and cultural belief.
What about language-based classifiers? Artificial intelligence, though increasingly used, remains ineffective to identify hate content in several languages, except for in Hindi and English, as demonstrated by the whistleblowers.
For most of our history, the content review process has been very reactive and manual -- with people reporting content they have found problematic, and then our team reviewing that content. This approach has enabled us to remove a lot of harmful content, but it has major limits in that we can’t remove harmful content before people see it, or that people do not report.
Over the last few years, we have significantly ramped up our efforts to proactively enforce our policies using a combination of artificial intelligence doing the most repetitive work, and a much larger team of people focused on the more nuanced cases. The single most important improvement in enforcing our policies is using artificial intelligence to proactively report potentially problematic content to our team of reviewers, and in some cases to take action on the content automatically as well. In 2016, the vast majority of our content removals were based on what users reported to us. We knew we needed to do better and so we began building technology to identify potentially violating content without anyone flagging it to us.
When we began reporting our metrics on hate speech, only 23.6% of content we removed was detected proactively by our systems; the majority of what we removed was found by people. Now, that number is over 97%.
We’ve invested significantly in technology to find hate speech in various languages, including 4 Indian languages - Hindi, Urdu, Tamil and Bengali. We will continue to build on this work and add more languages.
A lot of the hate speech identified in India has been found to be anti-Muslim (Islamophobic) and anti-Dalit rhetoric, according to several reports. Is there any attempt to focus on clamping down on such content?
We do not allow hate speech on Facebook and regularly work with experts, non-profits, and stakeholders to help make sure Facebook is a safe place for everyone, recognizing anti-Muslim and anti-Dalit rhetoric can take different forms.
We’ve invested significantly in technology to find hate speech in various languages, including Hindi and Bengali. As a result, we’ve reduced the amount of hate speech that people see by half this year. Today, it’s down to 0.05 percent. Hate speech against marginalized groups, including Muslims, is on the rise globally. So, we are improving enforcement and are committed to updating our policies as hate speech evolves online.
Could you please also elaborate on how your budget allocation has changed: is still aimed primarily towards the US, or has it been altered following the allegations?
The allegations that we spend more towards curbing misinformation and hate speech in the US than the rest of the world is out of context and does not give a comprehensive view of our efforts. The majority of the resources we have in place to prevent misinformation on our platform focus on content originating outside the United States. For misinformation, one of the primary ways we combat it is through our third-party fact-checking program, which includes over 80 partners globally, reviewing content in over 60 languages. The majority of those partners review content originating from outside the United States.
In India, we have invested in building a strong network of third-party fact checking partners. Today, we have as many fact-checking partners in India as in the US - 10 fact-checking partners with the ability to fact-check in 11 Indian languages.
In addition, For India, we have people reviewing potentially problematic content in 20 Indian languages and we have hate speech classifiers (automated detection technology) in four Indian languages - Hindi, Bengali, Tamil and Urdu. Our Community Standards are translated in 49 languages globally including Hindi, Bengali and Urdu.
What does the company think is the future of communication and how does it plan to implement it?
In the metaverse, the physical and the digital worlds would converge and it would be a more immersive and engaging experience, creating a shared sense of space with another person and another place. We believe that social experiences will be at the heart of the metaverse. As a social technology company - it is the natural next chapter for us as we help in building ‘bridges’ from our apps on 2D screens into more immersive virtual experiences in the metaverse.
Are there any new initiatives the company is planning to undertake to achieve its vision?
The next generation of online social experiences will feel immersive and new, this is the natural progression for the internet and of our own work today. It’s the next evolution in a long line of social technologies. If you look at the investments we’ve made over the years -- in communities, creators, commerce and the next computing platforms, like Oculus - you can see this vision starting to come into focus. We have always been about building social connections and with Metaverse we will follow a similar path, starting off with more basic bridges and glimpses in our 2D apps, and eventually becoming central to the way we connect online.
