Can tech companies help fight U.S. Census misinformation?
The U.S. election isn’t the only 2020 event threatened to be disrupted by misinformation.
Reuters reported Wednesday that the U.S. Census Bureau has asked Google, Facebook and Twitter to “help it fend off fake news campaigns it fears could disrupt the upcoming 2020 count,” citing Census officials and others briefed on the plans.
The decennial census, the constitutionally mandated count of every resident in the United States, is used for political and economic purposes. It not only determines the number of House of Representatives seats each state gets, but is used to calculate how billions of dollars in federal money are distributed among the states. Any disruption of the count could have a big impact on the lives of Americans.
There is already some angst surrounding the 2020 Census because it’s the first one to be conducted largely online (will the systems be ready?) and because of the ongoing legal dispute over whether people can be asked whether they are citizens, as the Trump administration wants. The Supreme Court is expected to hear arguments on the question next month.
Add to that the potential for misinformation spread on social media, and the possibility of trouble grows.
Dipayan Ghosh, director of the Platform Accountability Project at the Harvard Kennedy School, contacted via email, said nefarious disinformation operators are likely to attempt to infiltrate American information and media networks in the coming months with misleading content that suggests citizens should not or cannot participate — precisely because doing so can hinder the full functioning of our national democratic process.
Which brings us to the companies that run those platforms: Could they help? Ghosh thinks they can.
“Companies can build artificial intelligence systems to counteract this kind of negative behavior by inferring signals that can suggest who the disinformation operators are and where the misleading content is, and proactively take that sort of content down,” he said.
He noted that the companies have already done this kind of work in the case of Russian and other foreign disinformation operations pertaining to our national elections. They’ve also removed domestic disinformation operators and white supremacists from their platforms.
“They can do the exact same thing in regard to disinformation pertaining to the census,” he said. ”The question will be whether and how they can be fully incentivized to take this problem seriously enough.”
- Instagram (yes, Instagram) is “teeming with conspiracy theories, viral misinformation, and extremist memes,” Taylor Lorenz wrote in The Atlantic — and it’s largely escaped the kind of scrutiny that reporters have paid to Facebook and Google. Craig Silverman of BuzzFeed News also found conspiracy accounts getting massive engagement, and, in 2017, Jonathan Albright of the Tow Center wrote about how Instagram had become a major source of political propaganda. Meanwhile, Bellingcat has a new story about how it located The Netherlands’ most-wanted criminal using Instagram.
- GoFundMe banned users who raised money to spread anti-vaccine conspiracy theories, The Daily Beast reported. The move made GoFundMe the latest internet platform to take an aggressive stance on anti-vaxxer misinformation, following Pinterest, Facebook and YouTube, which faced pressure from American lawmakers to do so.
- You’ve probably heard by now that Facebook this week purged accounts linked to Iran, Russia, Macedonia and Kosovo, citing “coordinated inauthentic behavior.” The company provides examples on its blog.
Every candidate running for president in the United States is getting an unclassified report on the major national security challenges facing the country. The report, wrote The Washington Post’s Shane Harris, has the feel of an “urgent primer” and comes in response to “the recent rise and abundance of fake news and foreign election interference.”
CNN’s Oliver Darcy writes that Twitter’s algorithm is amplifying extreme rhetoric, often posted by ”media or internet personalities who hold fringe views.”
- We’ve written plenty about efforts in India to combat misinformation in advance of national elections this month and next. India’s citizens are concerned, too. Pew Research Center this week has published a survey about how the Indians get their information and how they view misinformation. It found that 77% say they are very or somewhat concerned about people being exposed to false or incorrect information when they use their mobile phones.
…the future of news
Reuters is training its journalists how to spot deepfake videos online, Digiday reported. It even created its own deepfake to emphasize the point. FWIW: the IFCN did this over the summer and it was really hard.
Google is assembling a team of philosophers, engineers, and policy experts to “help it navigate the moral hazards presented by artificial intelligence without press scandals, employee protests, or legal trouble,” MIT’s Technology Review reported.
Would requiring tech platforms to put all of their ads in a publicly searchable archive cut down on political misinformation? Writing for The New York Times, Philip N. Howard said the idea would “create a record of all such misinformation campaigns that could be used to prevent them in the future.”
Fact vs. Fake is taking a break this week in preparation for International Fact-Checking Day on Tuesday, April 2! Launched by the IFCN in 2016, the day aims to raise awareness of the importance of fact-checking in the modern era of misinformation.
Take our quiz to test your knowledge of misinformation.
Polish your digital verification skills through our Hands-On Fact-Checking Short Course.
Visit the YouTube Learning channel’s guest picks playlist celebrating fact-checking and digital literacy.
Read our tip sheets and learn how to do everything from verifying social media videos to geolocating images.
Connect with fact-checkers through the new, interactive EduCheckMap.
Participate in a Reddit AMA with PolitiFact at noon EST on April 2.
Watch NBC’s “Stay Tuned” daily news show on Snapchat as it features a special guest, “NBC Nightly News” anchor Lester Holt – the first MediaWise Ambassador – sharing MediaWise fact-checking tips with host Savannah Sellers.
Join the conversation by using #factcheckingday.
Amid all the fact-sorting involving the report of Special Counsel Robert Mueller this week, it was particularly important to parse the collusion/obstruction equation. PolitiFact did just that.
According to the letter to Congress from Attorney General William Barr, the special counsel did not establish that President Trump or his aides colluded with Russia to win the 2016 election, and he did not draw a conclusion as to whether the president obstructed justice. He left that to Barr, who declined to.
The attorney general noted that the lack of an underlying crime meant there was nothing to obstruct, though he acknowledged that wasn’t the sole determinant. The conclusion set off a whole debate about whether an underlying crime is needed to prove obstruction.
PolitiFact checked with 11 experts on the question, who generally agreed that obstruction of a non-crime is possible. They cited high-profile cases — those of former Dick Cheney aide Scooter Libby and TV personality Martha Stewart. But they said the precedents there included some nuances, which the fact-checker laid out.
What we liked: This was a good explainer of a complicated topic and, as such, did not include one of PolitiFact’s signature Truth-o-Meter ratings. It also included historical context, drawing parallels to Watergate, while also pointing out differences.
The fact-checker, Louis Jacobson, noted in the piece that Trump ended up pardoning Libby. One fun fact that he might have included: Trump thought Stewart also got a raw deal, and considered pardoning her, too.
- The Associated Press wrote about how Facebook is leveraging its fact-checking partners to cut down on misinformation about the Indian election.
- The reappearance of a longstanding online hoax warning of attempted kidnappings of young women or children has sparked a series of vigilante attacks on Roma living in France, the Guardian reported. The report said some 50 people armed with sticks and knives attacked Roma people and set fire to their parked vans.
- Google is releasing the beta version of its fact-checking search engine early next week.
- A “Fox & Friends” reporter cited a tweet from user called QAnon76 praising an executive order from President Trump on campus free speech, the Daily Beast reported, noting the user’s 160,000 followers.
- This may be a first. The Washington National Cathedral fact-checked the president on funeral costs for the late Sen. John McCain (R-Ariz.).
- Florida State University next week will host an event on truth and misinformation.
- From AFP: No, this is not a video of the man accused of killing 50 people in New Zealand mosques being beaten in prison.
- The IFCN’s new associate director, Cristina Tardáguila of Agência Lupa, spoke to CheckNews about how it tripled its fact-checking audience.
- PolitiFact has partnered with Telemundo to bring fact checks to millions of Spanish-speaking Americans ahead of the 2020 presidential election.
- Metafact, the new crowdsourced health fact-checking site we wrote about two weeks ago, has received full funding on Kickstarter.