Since its inception, the way content has been shared on Snapchat has made the platform less susceptible to the spread of fake news. This is because, historically, it has simply been a messaging app. The Discover feature is overseen by human moderators, so the chances of widespread misinformation are slim. Further, Snapchat employees are quick to analyze viral events or stories and fact-check them without letting them get out of hand on the platform.
Twitter has clearly stated its stance on misinformation :. Journalists, experts and engaged citizens Tweet side-by-side correcting and challenging public discourse in seconds.
In short, the platform does not seek to combat misinformation directly. It does, however, take action against spammy or manipulative behaviors, particularly when it comes to bots.
Indeed, millions of accounts have been suspended for this reason. Throughout the past decade, misinformation has flourished on social media. While information on users — including their demographics and internet behavior — is intended to be used by marketers, it can also be leveraged by those looking to spread misinformation.
Using this data, fake news can be designed to appeal to a core audience to encourage authentic engagement. From there, the content can gain exposure rapidly, potentially going viral. In this way, social media enhances the reach of those responsible for misinformation. Each social media network contributes to misinformation in different ways, impacting how people handle issues related to politics, health, and more.
Many have argued that fake news on Facebook had a dramatic impact on the presidential election. In fact, an article published by The Atlantic states that the social media giant has fundamentally altered the nature of American Democracy. The platform is a major source of political information for millions of Americans, and fake news can have a powerful effect on elections. For example, one image that gained popularity in late was that of President Trump helping flood victims in Texas after Hurricane Harvey.
However, the image of Trump was fake. This image gave a false impression that impacted public perception. Digitally manipulated images, in addition to fake news presented in Instagram Stories, are powerful ways to spread misinformation. Overall, fake news is less of a concern on Snapchat than on other social media platforms discussed here. Nevertheless, this has changed slightly with the addition of the Discover feature, which could hypothetically be used to propagate misinformation.
This feature uses machine learning to suggest content that users may enjoy based on their viewing history. Misinformation on Twitter has proven difficult to prevent. Further, while Twitter officials have sworn to take action against bots, research indicates that they are not able to keep up with the flood of junk accounts. How can you recognize misinformation on social media? It often has a clear bias, and it may attempt to inspire anger or other strong feelings from the reader.
Such content may come from a news source that is completely unfamiliar, and the news itself may be downright nonsensical. Do either have an established reputation?
Are they known as trustworthy sources? If not, do they cite their sources — and are those reputable? Fake news often uses fake author names and bogus sources. You may notice suspicious details. Be kind. Bennett, The Light in the Heart. But I never believe me. Hinton, The Outsiders. If people all over the world Browse By Tag.
Welcome back. And for the record, there have been wildfires in Canada this summer. PolitiFact also rated these claims False. In August we fact-checked a widely shared Facebook post that alleged the president had denied fire aid to California but then helped Russia fight wildfires in Siberia. We rated that False when it popped up, but it is circulating again. Some of these really outlandish claims seem to have nine lives and just appear over and over again. Just to recap: President Trump has repeatedly threatened to withhold federal fire assistance from California and he made a similar threat again last month.
But he has always approved that aid. On Russia, it is true that Trump said last year that he offered Russia help with its fires in Siberia. We found no evidence that he actually followed through.
It's no secret that the posts and information on Facebook, whether fact or fiction, can influence what people believe and how they act.
But machine-learning models do not work that way. It must be trained on thousands, often even millions, of examples of a new type of content before learning to filter it out. Even then, users can quickly learn to outwit the model by doing things like changing the wording of a post or replacing incendiary phrases with euphemisms, making their message illegible to the AI while still obvious to a human.
This is why new conspiracy theories can rapidly spiral out of control, and partly why, even after such content is banned, forms of it can persist on the platform. Schroepfer and his more than engineering specialists create A. Meanwhile, the algorithms that recommend this content still work to maximize engagement. This means every toxic post that escapes the content-moderation filters will continue to be pushed higher up the news feed and promoted to reach a larger audience. And even after it merged into the bigger Responsible AI team, it was never mandated to work on content-recommendation systems that might limit the spread of misinformation.
But Schroepfer told me precisely the opposite in an earlier interview. I had asked him why he had created a centralized Responsible AI team instead of directing existing teams to make progress on the issue. If everyone gets the recommendation, does that mean it was fair? It allows engineers to measure the accuracy of machine-learning models for different user groups. One of the thornier problems with making algorithms fair is that there are different definitions of fairness , which can be mutually incompatible.
Fairness Flow lists four definitions that engineers can use according to which suits their purpose best, such as whether a speech-recognition model recognizes all accents with equal accuracy or with a minimum threshold of accuracy. But testing algorithms for fairness is still largely optional at Facebook.
Pay incentives are still tied to engagement and growth metrics. This last problem came to the fore when the company had to deal with allegations of anti-conservative bias. These include various dimensions submitted by users or estimated by machine-learning models, such as race, political and religious leanings, socioeconomic class, and level of education. The Fairness Flow documentation, which the Responsible AI team wrote later, includes a case study on how to use the tool in such a situation.
If conservatives are posting a greater fraction of misinformation, as judged by public consensus, then the model should flag a greater fraction of conservative content. If liberals are posting more misinformation, it should flag their content more often too.
When a model did so, they would stop its deployment and demand a change. Once, they blocked a medical-misinformation detector that had noticeably reduced the reach of anti-vaccine campaigns, the former researcher told me. They told the researchers that the model could not be deployed until the team fixed this discrepancy. But that effectively made the model meaningless.
This happened countless other times—and not just for content moderation. And ahead of the election, Facebook policy executives used this excuse, according to the New York Times , to veto or weaken several proposals that would have reduced the spread of hateful and damaging content.
Many of these incidents happened before Fairness Flow was adopted. I wanted to know how the storming of Congress had affected his thinking and the direction of his work. I asked him the question again. His Facebook Portal camera, which uses computer-vision algorithms to track the speaker, began to slowly zoom in on his face as he grew still. He told me it was the job of other teams though none, as I confirmed, have been mandated to work on that task.
I pressed him one more time. Honest to God. Corrections: We amended a line that suggested that Joel Kaplan, Facebook's vice president of global policy, had used Fairness Flow. He has not. But members of his team have used the notion of fairness to request the retraining of misinformation models in ways that directly contradict Responsible AI's guidelines.
0コメント