NewsGuard solves the wrong problem for today’s news landscape

This week NewsGuard launched to much fanfare as it heralds the end of fake news with “nutrition labels” for 7500 news sites. Unfortunately the effort misses the real issues that plague the news today — news bias and variable quality journalism.

Fake news — more about bias than falsehoods

Here’s what NewsGuard aims to do with its “nutrition labels”:

The labels will explain the history of the site, what it attempts to cover, who owns it, who edits it, and make transparent other relevant factors, such as financing, notable awards or missteps, whether the publisher participates in programs such as the Trust Project…

This is good. Knowing who finances any publication is important in understanding what agenda they may have.

But then NewsGuard says this:

We are not going to make granular judgments about political leanings…

This is where they miss the forest for the trees.

A recent survey from Knight Foundation (pg. 27) shows that people consider news bias, negative framing, and shoddy reporting — all as “fake news”. And these problems are far more ubiquitous than outright falsehoods.

Indeed, the issue of news bias and framing is what is driving a wedge in our society. Take a look at the picture below from Pew Research. The divide in our society is not because we as humans changed our attitudes so dramatically in ~10 years (the prior 10 years saw no change). Rather, what has changed is that most of us get our news from social media (p. 12) which gives us sources that confirm what we want to believe rather than what the truth is.

Bridging the above divide begins with helping readers see different viewpoints on issues they hold dear. All issues are more complex than a single headline or article may suggest. Helping people find and read multiple high quality perspectives so they can zero in on the truth is far more needed than labels about a publication’s history.

Articles not publications

The point about “not making granular judgements” is another shortcoming of NewsGuard.

NewsGuard gives a single rating to an entire news site. But a single rating to describe the work of every journalist at a publication is not very useful.

As an example, I’m guessing that NewsGuard will rate Breitbart as “Yellow” or maybe even “Red”, and Washington Post as “Green”. But Breitbart is a top-50 news site in the US, ahead of Washington Post. How does the label help me understand why the site is so popular?

The reality is that Breitbart and Washington Post, like every news site in the world, has bias and variable quality in its articles. Some articles are excellent with authors taking great pains to cite several credible and diverse sources (including those that disagree with their thesis) while other articles are poorly sourced or essentially propaganda. Helping readers differentiate good quality articles from poor ones is more important than branding entire sites as “green” or “yellow.”

Humans and machines must work together

NewsGuard says humans are better than algorithms in evaluating news quality:

Our goal is to help solve this problem now by using human beings — trained, experienced journalists — who will operate under a transparent, accountable process to apply basic common sense to a growing scourge that clearly cannot be solved by algorithms...

Certainly Google and Facebook’s algorithms have failed to identify fake news. But to say that algorithms are incapable of solving this issue and resorting to humans alone is to ignore the long list of cognitive biases and inconsistencies that humans have.

The solution to identifying quality news is likely a combination of the two approaches. Algorithms are generally consistent but often have unintentional biases. Human oversight helps to catch and mend these shortcomings and improve the algorithm over time. But neither approach on its own is the answer.

Step in the right direction

Even with the above shortcomings I commend NewsGuard for what they’re doing. They will provide transparency in funding sources behind publications and that is a key piece of the puzzle. But it is not enough to solve the problem of fake news, or more importantly, a divided and dysfunctional society.

Disclosure: I have a dog in this fight and am not neutral myself. I am the founder of CivikOwl which helps people figure out what news they can trust.