I'm not sure if any studies accurately measure the subject content, and it also seems like it'd be easy to get any result you want by tweaking the experiment parameters.

I can tell you that from my own life experience and what I can recall, it has certainly felt like a shift in content. All the anti-white stuff, the social justice stuff, the pro-censorship stuff, etc. It existed before in tiny bubbles like Tumblr, but it was during the mid 2010s that the major news sources started adopting those viewpoints too.

They don't even show it as a percentage of articles.

It's suspicious to me that they went for the hockey stick curve to demonstrate their complaint rather than adjusted for the number of articles that they are sampling from.