Throughout 2017, we’ve spent a lot of time thinking, researching, writing, and talking about how disinformation flows online, about filter bubbles and fact-checking, transparency for news organizations, and other important pieces of the puzzle about how the internet has changed the way so many of us consume and understand information.

But there’s been far too little attention paid to an older form of communication that still has deep influence in our democracy: that old-fashioned thing known as the television, specifically, TV news.

Why? Not because TV news networks, including Fox, MSNBC, and CNN, don’t have influence, but rather because television is difficult to study. Informed voices have urged Facebook to release its data to fact-checkers and others working to improve the quality of news shared online. But TV news content remains both opaque and ephemeral. TV news networks make their content available online generally, but viewers are at their mercy when searching for particular clips, sharing such information elsewhere, or providing structured datasets to help inform research. If a network goes defunct, like Al Jazeera America did, we don’t have any guarantee that material will be preserved.

Enter the Internet Archive, whose mission is to provide universal access to all knowledge. Most journalists know us for the Wayback Machine, which has preserved more than 308 billion webpages online. But the Internet Archive is also home to the TV News Archive, whose collection includes more than 1.4 million TV news shows, searchable by closed captions. We are working hard with partner organizations, with journalists, and with researchers, from Duke Reporters Lab to PolitiFact, from Stanford University to startups like Matroid and Joostware, to turn our archives into data. We are applying machine learning to generate structured data in increasingly sophisticated ways, so that ultimately it will be possible not just to search captions for TV news, but also faces, talking points, identify who is speaking, and more.

For example, in 2016 we launched the Political TV Ad Archive, which used an open source audio fingerprinting tool we called the Duplitron to track political ad airings across key media markets. We fed this information to our fact-checking and journalism partners, who mined it to report on the 2016 elections.

In 2017, we developed the Trumpcongressional, and executive branch archives, curated collections of clips by key political and administration figures that can searched by keywords and phrases. We also created Face-o-Matic, which tracks the faces as shown on cable TV news of President Donald Trump and the four top congressional leaders, and Third Eye, which extracts chyrons, or the lower thirds of TV screens, and turns them into downloadable data ready for analysis. The New York Times editorial page, for example, used Third Eye to track how cable news networks differed in their coverage of key indictments in the Russia investigation by special counsel Robert Mueller.

In 2018, we plan to take even greater strides in helping us understand ourselves through TV news. “Who Said What,” by Joostware and “Contextubot,” by Bad Idea Factory, two of the winners of the Knight Foundation’s (in partnership with the Democracy Fund and the Rita Allen Foundation) call for projects to combat misinformation, rely on the TV News Archive to fuel their projects. We’re working with the Duke Reporters Lab on its Tech & Check project to help automate the workflow for fact-checkers. And we’re developing new partnerships with institutions like Stanford University to develop new ways to turn our TV News Archive into data.

We’re talking to media literacy educators about deploying TV News Archive materials into curricula. And in this age of media manipulation, we are exploring ways that we can authenticate TV news clips, so the viewer knows they have not been altered. Finally, we’re expanding our collection of TV beyond national borders, because understanding how others in the world view us, as well as how we view them, will be crucial in the years to come.

Even in 2018, in the era of tablets and phones and Twitter and Facebook and Instagram, TV still affects us all. Knowing our TV is crucial to understanding ourselves.

Nancy Watzman is managing editor of the Internet Archive’s TV News Archive, and is director of strategic initiatives fo Dot Connector Studio. This piece appeared in the series, “Predictions for Journalism 2018,” in NiemanLab.