Machines + Media 2018: New applications of data science and technology in media and journalism

The media industry is on the verge of disruption. While the basic tenets of journalism are to present a balanced view based in facts, every aspect of storytelling and distribution is affected by technology. Even though the industry has more tools than ever to write, report, produce, and distribute stories, fake news sometimes prevails because that’s what people want to believe. Journalism’s best defense against fake news is investments in technology and innovation.

“There’s a sense of opportunity and potential for innovation and also a sense of trepidation about what some of these developments mean,” said Justin Hendrix, executive director at NYC Media Lab. “Like any technology, we’re playing with fire – fire is an incredibly valuable resource for humanity, but it’s also very dangerous.”

At the Bloomberg-sponsored Machines + Media 2018 conference last week, organized by NYC Media Lab and hosted by Bloomberg at its Global Headquarters in New York City, industry executives, academics and politicians discussed the business, technology, and innovation in the media sector, along with privacy and ethical concerns.

There’s good reason to believe that the future of news will be more automated, more paid for and eventually less fake. “Most of the impact of technology is on the distribution side,” said John Micklethwait, Bloomberg editor-in-chief. “The idea that you could reach just about everybody almost immediately would have seemed unbelievable not that long ago, but that’s the world that we’ve got.”

Technology has sped up the news cycle. Today, news is incredibly valuable for a short period of time – about 10 to 15 seconds for some markets – before the debate quickly turns to why and what’s next. When machines, through artificial intelligence and machine learning that can automate tasks, capture basic facts for events that are naturally repeatable, journalists can instead focus on the why. People aren’t replaced, but rather become specialized in the creative tasks that machines can’t perform. “Let people focus on what they’re best at, then the ROI on using artificial intelligence in any industry will soar massively,” said Drew Silverstein, CEO of Amper Music.

The distribution of news has changed as well. News has evolved from something you get and consume to something that follows you around. Technologies like blockchain, augmented reality and natural language processing can help journalists disseminate information in more engaging and unique ways, but content has to be delivered at the right time. Personalization affects what people read – sometimes people read an article because it’s true and other times, because it’s appealing.

“We’d like to think that people are interested in the facts, but many people are interested in what interests them,” said Amanda Stent, NLP Architect with the Office of the CTO at Bloomberg. “The challenge for media organizations is to present people with enough information to bring us back into a common space by giving people the information they need to know right now, regardless of whether it’s something they personally chose.”

Getting people to that common space becomes difficult when technology is also used to create and distribute content containing little to no facts or altered realities, no matter the medium. “There used to be a cost associated with publishing content online,” said Mor Naaman, associate professor of information science at Cornell Tech. “Think about the 1990s and the early web, having a well-designed website with a decent URL meant that you are a credible organization. Today, it’s so easy to produce content, and there’s so little risk. In fact, there is a reward for producing fake content.”

Just like technology can be used to alter the truth, technology can also protect the truth. Blockchain, for example, could be used to track stories as they travel the Internet so that readers can verify where that story was originally published. “Blockchain is decentralized, and the crypto key goes along with the content no matter how it’s shared,” said Clay Eltzroth, product manager for Bloomberg News. Being able to follow content to the original source because of a tag can help to reinforce trust in an outlet and the news the outlet produces.

There are many more applications of technology that have yet to be explored. In any industry undergoing transformation, startups are often the best place to find these new applications. During Machines + Media, six companies presented their solutions to some of the challenges facing the media today:

  • AI is shaping the future of storytelling and automated stories, but these tools often miss global voices and cultural data, and the storytelling agency IVOW adds that diversity to AI by telling stories about cultural traditions.
  • When much of the world’s population and land don’t have addresses, location isn’t a solved problem. Foam provides proof of location using blockchain protocols.
  • There has been a big shift from broadcast TV to being able to watch TV anytime, anywhere. With so much content and variety, things start to look the same. Tagasauris improves how people interact with video by creating micro-moments, or video snacks, of interesting hot spots in a video. These are embedded with contextual hyperlinks to create story-like experiences that can be adequately shared on the web.
  • Machine learning is a powerful tool, but this fast-changing industry lacks the most fundamental infrastructure – when an engineer leaves a team, their knowledge goes with them. To allow companies to regain control over their machine learning experiments, Comet provides a mechanism for teams to collaborate and for that knowledge to stay within a company.
  • Media companies set the global standard to curate, acquire and create the best content, but they must also be able to distinguish between real and fake news, and catalog, understand and search all the content. Algorithmia automates the DevOps process for machine learning so data scientists can focus more on ideas.
  • Making video content searchable, understandable and discoverable so users can find what they’re looking for is important. Vidrovr extracts content from videos to bring powerful contextual search solutions to help keep users on a media outlet’s website.

“After a year of torrid discovery about the media and technology ecosystem,” says Hendrix, “there are clearly many reasons to be optimistic about the future of the industry.”