Combining AI and an editorial team can prevent "fake news" and other unreliable information from influencing business decisions.
Not all content is created equal. Moreover, not all content can be trusted. New York Times followers received a sharp reminder of this axiom on Sunday, January 22, when OurMine, a hacker “security” group, infiltrated one of the media organization’s Twitter accounts. OurMine tweeted a fake news brief through the New York Times' video account (@nytvideo) claiming that Russia had launched a missile attack against the United States. The source was supposedly a “leaked statement” from Russian President Vladimir Putin. The Times quickly removed OurMine’s fake tweet: “We deleted a series of tweets published from this account earlier today without our authorization,” they reported via Twitter. Unfortunately, not everyone seems to realize that not all content can be trusted.
“Fake news” is believable partly because so much information is presented in a uniform style. The acceptance of social media posts, for example, is strengthened by Twitter’s consistent format, which the receiver often instantly recognizes and accepts. Since the viral spread of tweets is immediate, many people can be fooled very quickly. Fake news sent by hackers through business-to-business channels can have a similar influence. Its consequences potentially damage both the news distributor (possibly a corporate website), whose reputation suffers, and the organization receiving it, which may disregard legitimate information from the same source. Fake news can harm client relationships and impede the growth of further accounts. The same can be said to a lesser degree of badly organized “honest” content, loosely gathered from the internet, but not properly vetted.
The proliferation of fake news has made data management and security hot issues, particularly for information professionals. In 2016, the theft of sensitive data was highly publicized for its damaging effects on the reputations and finances of private companies. Corporate officials are now addressing this issue with an increased focus on data security. This focus is blurred by the never-ending flow of information, which threatens to overwhelm the capacity of companies to administer and distribute it. Securing data is partly an issue of data management, choosing which information needs better protecting, which is inconsequential, and which is redundant, out-of-date, and trivial (R.O.T.). The status of content is not static; as it ages, it generally becomes less important, at least to the company holding it. Typically up to 70% of enterprise data is R.O.T.
When R.O.T. content is truly obsolete, its cost and inconvenience can be considerable. Microsoft.com reportedly had over 10 million pages of content available online at one point, of which nearly three million had never been accessed. Finding relevance in such a huge warehouse of information is often a matter of luck. Vast stores of R.O.T. content can’t linger without creating unforeseen problems and future risks. The information archive released by Edward Snowden had little relevance to Booz Allen, the company holding it, but the ramifications of its release for the U.S. government were enormous. Companies wishing to manage information securely must eventually decide what to do with these archives, which to delete, which to organize, and how to address the issue of redundant or bad content versus good.
When it comes to content, separating the wheat from the chaff calls for curation. Some curation is more easily accomplished by using Machine Learning technology to scrub information. The International Institute for Analytics (IIA) predicted in 2016 that technological forms of data curation would grow increasingly important over time. “The new data analytic tools work from the bottom up,” wrote the IIA, “leveraging machine learning to curate and clean the data.”
Since volumes of data are increasing, companies find it increasingly difficult to address issues of data security without using AI. The ability to sort through huge amounts of unstructured data is a strength of Machine Learning. Because so much content is now distributed across different platforms and networks, tracking it can take considerable effort. Machine learning can detect patterns in the data and so recognize the rules by which it is organized. Its approach can eliminate the necessity of reclassifying data manually as further information is introduced. Such data curation makes it easier for Corporate Security Officers to observe regulations governing the security of information. CSOs can more easily track what information is being used and with whom it is being shared, and so decide what information should be deleted, archived, or better protected.
Certain sophisticated forms of social network curation are not without their own problems for those who implement it. Researchers at Oxford University have argued against employing a “recommendation algorithm” to sort data and predict what a user will like best among a group of given items. Such an algorithm can predict what a user will want at any moment based on the user’s prior history of selections, as measured by the network that presents the information. A feedback mechanism optimizes the user’s interaction with the network. The researchers argue that methods of targeting specific users at specific times are forms of manipulation, pointing to the 2016 US Elections, when fake news stories were used as a way to increase “hits” to websites and thus generate advertising revenue. The team argues that using these curator algorithms may take advantage of the emotional state of the user.
But these algorithms are tools; human editorial expertise must also play a significant part in content curation for it to be effective. Business and government organizations would be well-advised to stay abreast of the ways that humans and AI together can ensure quality information. Each aspect is crucial in combating the effects of fake news and hacking that spreads it. Now that the New York Times has been thrust into the middle of that combat, the battle for verified content is likely to intensify, with the only sure loser being the trusting reader.
Tame the Ever-Increasing Flow of Information
InfoDesk has created the world’s smartest platform for managing and sharing information. With our comprehensive solutions, you can bring all your information together, filter and select relevant content, and deliver the right intelligence to the right people. InfoDesk has been providing actionable intelligence to multinational corporations, government agencies and other organizations since 1999. InfoDesk is based in New York with offices in London, Washington, DC and India. Learn more about InfoDesk.