Hillary Clinton has personally ordered several assassinations; Donald Trump once referred to Republican voters as “the dumbest…in the country”; and literally everyone is a Russian agent.

FAKE NEWS is everywhere.

The sheer quantity of misinformation on the internet is as evident as the need to do something about it. Luckily Google, along with its sister think-tank Jigsaw, is attempting just that.

Google first started tagging certain bits of content as fact-checked in October last year. Certain news articles earned the “Full Fact” label, helping users quickly judge the trustworthiness of the source at a glance.

At the time, the label only applied to certain articles in Google’s News section, and only in the UK and the US. Now, fact-checking has gone global. It is now available in all countries, in all languages and, importantly, in basic search results.

Here’s a much touted example of the new fact-checking label in action:

Clearly displayed is the main claim in question, its source, and an assessment of it’s truth, along with the name of the publisher who made the assessment.

This last bit is important – Google are making sure to distance themselves from the judgement itself, making it clear that “these fact checks are not Google’s and are presented so people can make more informed judgements.”

HOW IT WORKS

In order for a claim to be shown as verified/falsified, the publisher in question must either use the ClaimReview markup from schema.org, or use the Share the Facts widget, developed by Jigsaw and the Duke Reporter’s Lab.

The ClaimReview markup is included in the appropriate page by its publisher, specifying the relevant details – i.e. a summary of the claim in question, who made it, and, of course, whether or not it is true. Multiple fact-checks can be applied to multiple discrete claims on any one page. In such a case, you’ll see a fact-check carousel on the SERP if using a mobile browser.

(For more on how schema markup and structured data work, read this article)

More information about the Share the Facts widget can be found here, but here’s what it looks like in action on a page:

And here’s the same page displayed in Google:

WHO FACT CHECKS THE FACT CHECKERS?

Whichever signalling method is chosen, the content in question must adhere to Google’s News Publisher guidelines relating to “standards for accountability and transparency, readability or proper site representation”.

For example: “Analysis must be transparent about sources and methods, with citations and references to primary sources.”

Importantly, “fact checks are not guaranteed to be shown” merely due to their inclusion by the publisher.

Google tells us that “only publishers that are algorithmically determined to be an authoritative source of information will qualify for inclusion”.

Google has remained characteristically opaque with regard to how they actually judge the authoritativeness of a given publisher, beyond explaining that the final decision will be made by the all-powerful algorithm.

WHAT IF THERE’S DISAGREEMENT?

There will be times, Google acknowledges, when the truth of a claim is not a simple matter, and times when there is disagreement over its truth altogether. Even among sources Google has algorithmically deemed trustworthy.

There is room for this in the new system though. For example, as we can see in the snippet above from PolitiFact, the claim is rated as ‘half true’. Similarly, some claims will get a ‘mixed’ fact-check result from Snopes.

“There may be search result pages where different publishers checked the same claim and reached different conclusions” Google explains.

“Even though differing conclusions may be presented, we think it’s still helpful for people to understand the degree of consensus around a particular claim and have clear information on which sources agree.”

……

Currently the new fact check labelling system is in its early stages, and so we’ll have to wait and see how truly effective it is. For now, we know for certain that it will have no effect on the ranking of the pages in question, and we know that the criteria by which a publisher can be judged authoritative enough to actually perform the fact-check are tight. It is plausible that partisan, if not entirely malicious, publishers will find ways to falsely fact-check certain bits of content, and it remains to be seen how effective Google’s safeguards against this actually are.