“It’s unlikely that within the next two years, much less the next few months, Facebook can build up protections in every country. But it must start planning now for how it will exponentially scale up people, products and partnerships to handle so many elections at once in 2022 and 2024.”New York Times
Kenya is scheduled to have a General Election on 9 August 2022. Voters will elect the President, members of the National Assembly and Senate, county governors and members of the 47 county assemblies.
But there’s a real danger that it will be hit by social media hate speech.
Source: New York Times
I Worked at Facebook. It’s Not Ready for This Year’s Election Wave.
Jan. 29, 2022
By Katie Harbath
Ms. Harbath is the chief executive of Anchor Change, a company focused on issues at the intersection of tech and democracy, and director of technology and democracy at the International Republican Institute, a nonpartisan organization committed to advancing democracy and freedom. She is a former public policy director at Facebook.
The world is not ready for the coming electoral tsunami. Neither is Facebook. With so many elections on the horizon — France, Kenya, Australia, Brazil, the Philippines and the United States will hold elections this year — the conversation now should focus on how Facebook is preparing.
I know what it’s like to prepare for an election at Facebook. I worked there for 10 years, and from 2014 through the end of 2019, I led the company’s work across elections globally. It has poured more than $13 billion into building up its safety and security efforts in the United States since the 2016 elections, when the platform was too slow to recognize how its products could be weaponized to spread misinformation.
Responsible election plans cannot be spun up in days or weeks. It takes time not only to organize internally but also to make meaningful and necessary connections with the communities around the world working to secure elections. Facebook must begin serious, concerted, well-funded efforts today.
For some of the elections happening in the first half of this year, Facebook is cutting it close. But there’s still time for Facebook to commit to a publicly available road map that outlines how it plans to build up its resources to fight misinformation and hate speech around the world. Algorithms that find hate speech and election-related content; labels that give people more context, like those in the United States applied to content that questioned the election results; and efforts to get people accurate information about where, when and how to vote should all be a part of the baseline protections Facebook deploys across the globe. On top of these technical protections, it needs people with country-specific language and culture expertise to make tough decisions about speech or behavior that might violate the platform’s rules.
I’m proud of the progress the company made in bringing more transparency to political and issue ads, developing civil society partnerships and taking down influence operations. None of that progress happened spontaneously. To combat the Internet Research Agency, a Russian troll farm that exposed 126 million Americans to its content before and after the 2016 elections, for example, Facebook needed new policies, new expertise and a revamped team at the platform dedicated to these issues. Because of those innovations, the company was able to take down 52 influence networks in 2021.
Facebook couldn’t do this work alone. Partnerships with organizations such as the Atlantic Council, the National Democratic Institute, the International Republican Institute and many others were crucial.
But even then, providing the technical infrastructure to combat misinformation is only half the battle. Facebook faced scrutiny again in 2020 and 2021 for how it handled everything from President Donald Trump’s Facebook account to false election fraud claims and Jan. 6. Many of the conversations I had at the time revolved around balancing the right to free speech with the harm that speech could cause someone.
This is one of the central dilemmas companies like Facebook grapple with. What is the right call for company administrators when a sitting president of the United States violates their platform’s community standards, even as they believe that people should be able to hear what he has to say? When are people exercising their right to organize and protest against their government, as opposed to preparing for a violent insurrection?
Similar issues come up in other countries. Last year the Russian government pressured Apple and Google to remove an app created by allies of Aleksei Navalny, an opponent of President Vladimir Putin’s. Refusing the government would have put their employees in Russia at risk. Complying would go against free-expression standards. The companies chose to protect their employees.
These are the kinds of difficult questions that crop up in every country, but Facebook also needs country-specific monitoring. Human expertise is the only way to truly understand how heated discussions are shifting in real time and to be sensitive to linguistic and cultural nuances. The word “dill” in Russian translates to “ukrop,” for example, which has been used as a slur against Ukrainians. Some Ukrainians, however, reclaimed the word and even named a political party after it. A global framework that fails to account for these kinds of situations or that is overly reliant on technology to address them is not prepared to confront the reality of our complex world.
Facebook has invested billions in this kind of work. But a majority of its investment for classifying misinformation, for example, has focused on the United States, even though daily active users in other countries make up the vast majority of the user base. And it’s not clear which efforts Facebook will extend from U.S. elections to those in other countries. It’s unlikely that within the next two years, much less the next few months, Facebook can build up protections in every country. But it must start planning now for how it will exponentially scale up people, products and partnerships to handle so many elections at once in 2022 and 2024.
It should be transparent about how it will determine what to build in each country. In 2019, Facebook had more than 500 full-time employees and 30,000 people working on safety and security overall. Even with that amount of human talent, it could cover the national elections in only three major countries at once. At least that many people were needed for the United States in 2020. In two years, people in the United States, India, Indonesia, Ukraine, Taiwan, Mexico and Britain are to go to the polls in national elections. Facebook will need to consider hiring at least 1,000 more full-time employees to be ready for the next big election cycle. If the company is cutting it close for 2022, it has just enough time to be really ready for 2024.
These problems are not ones that Facebook can fix on its own. Its parent, Meta, is a private company but one with tremendous influence on society and democratic discourse. Facebook needs to continue to recognize the responsibility it has to protect elections around the world and invest accordingly. Governments, civil society and the public should hold it accountable for doing so.
Katie Harbath is the chief executive of Anchor Change, a company focused on issues at the intersection of tech and democracy, and director of technology and democracy at the International Republican Institute, a nonpartisan organization committed to advancing democracy and freedom. She formerly worked at Facebook, where she helped lead its work on elections.