"Labeling is more art than science. And short of labeling all news media, their ownership, and funding, the classifications are imperfect - but more information about how they are made and what, if any, impact they have would help experts refine them."
This statement was originally published on cpj.org on 12 August 2020.
By Courtney C. Radsch/CPJ Advocacy Director
Twitter announced last week that it would start labeling some accounts run by media outlets and their top editors as “state-affiliated,” a descriptor intended to improve transparency about the source of information being shared on the platform.
Since disinformation became a flash point in the debate over content moderation on social media, distinguishing propaganda from public service news has become a priority for companies that operate global platforms – especially when editorially independent outlets receive government funding. Facebook and Google, which owns YouTube, apply their own versions of Twitter’s label. Twitter has also said it will not accept advertising from state-controlled media, while Facebook blocks state media accounts from advertising to U.S. audiences, citing concerns about political information being manipulated ahead of elections in November. YouTube representatives have told CPJ the company uses the same label to identify state-funded outlets and their advertising.
The classification efforts appear timed to stave off government interference. When top Facebook and Google executives appeared before a congressional antitrust committee in July, they had to answer questions about how their platforms address disinformation. A UK Parliamentary inquiry has separately called for an independent regulator to impose a compulsory code of ethics on companies perceived to enable the spread of “fake news,” with the power to bring legal action against them for violations. As policymakers continue to probe the dominance of the social media platforms – and their role in facilitating the global spread of disinformation that influences decision-making about elections and other democratic institutions – companies may be hoping that self-regulation helps preempt such controls.
Labeling state media is fraught. In separate investigations, CPJ and ProPublica have found YouTube’s labeling to be inconsistent. Several observers and people working at outlets that have been subject to such designations say the process lacks transparency, according to CPJ interviews and public statements. Some said they perceive the labels as derogatory and subjective; others welcomed them, noting their potential to improve transparency for users regarding where information comes from. However, many agreed that categories with more nuance, applied to a wider array of media organizations, would make labels more effective.
Determining the level of state interference in a given media outlet requires considerable expertise, according to several journalists and experts. Many told CPJ they agree the platforms have a role in addressing propaganda which stokes disinformation. Yet neither they, nor the companies responsible for the policies, could point to evidence of the impact of labeling. It’s not clear that identifying such outlets on social media improves media literacy or reduces the reach of propaganda – or whether it changes the way readers engage with labeled content. In the meantime, however, the companies are adopting definitions with the potential to reduce the visibility of select media outlets, their ability to advertise, and how audiences encounter their work on social media.
“I think most of us are in favor of helping users understand where their news is coming from,” said Matthew Baise, the director of digital strategy for the U.S. Congress-funded broadcaster Voice of America (VOA), who oversees its social media accounts in 47 languages.
“I’m all in favor of labeling,” said Anya Schiffrin, a former journalist who directs technology, media, and communications studies at Columbia University’s School of International and Public Affairs. “If you don’t want regulation then you have to believe in consumer choice, and labeling is part of promoting consumer choice.”
Twitter’s August 6 announcement says the company will apply the policy to accounts run by news media that they determine are affiliated with the state, as well as their senior staff, starting with those from the so-called P5 countries – China, France, Russia, the U.K., and the U.S. – that hold permanent positions on the U.N. Security Council. State-financed media that enjoy editorial independence, like the BBC in the U.K. and U.S. National Public Radio (NPR), would not receive the label, the announcement said. Accounts run by government spokespeople, like foreign ministers and ambassadors, will also be labeled.
Labeled accounts and their content will no longer be amplified through the platform’s recommendation systems, Nick Pickles, Twitter’s public policy strategy director, told CPJ – meaning that readers may be less likely to see their posts.
Twitter selected the initial target countries because of their “outsized influence on the global conversation,” Pickles told CPJ. These states also fund and operate media outlets in scores of languages and jurisdictions, making them harder to assess. In 2019, CPJ documented an “information war” that social media platforms had identified involving accounts originating in mainland China, where the platforms are generally censored. The “significant state-backed information operation,” as Twitter described it, sought to discredit pro-democracy protesters in Hong Kong. According to Pickles, journalists working for Chinese state media were running ads and tweeting about the protests without their affiliation being clearly visible in posts, including some that went viral.
Google, which began identifying “state-funded” outlets on YouTube in 2018, now labels content from both state and publicly-funded news outlets in 22 countries, primarily in Europe and the Asia-Pacific region. Facebook announced its policy in 2019 but delayed implementation, in part due to lobbying by resistant media outlets. After consulting with 65 experts and organizations, including CPJ, Facebook said in June this year that it would begin identifying media content and ads from outlets as “state-controlled” if they were “wholly or partially under the editorial control of their government,” excluding public service media.
“The purpose of the labels is to provide context on the video that people are viewing,” a Google official told CPJ in June. The aim was “not to dissuade people from viewing content, but to understand that the content they are viewing is coming from a government channel,” the official said, declining to be identified by name per company policy.
