The Tara Reade/Joe Biden Story is Testing the Limits of Big Tech Censorship

Their “equitable algorithms” are neither equitable nor algorithms

Background and Timeline

I’d prefer to start this blog with “You’ve probably heard about the recent allegations against Joe Biden…” but that might not be true. So instead I’ll start with a brief timeline of this important story:

• March 25: In an interview with Katie Halper, Former Biden staff assistant Tara Reade claims Biden sexually assaulted her in 1993.
• April 9: Reade files a police report in Washington DC, citing safety reasons and acknowledging that the statute of limitations has passed.
• April 24: The Intercept reports on the Aug. 11, 1993, episode of Larry King Live, where a caller speculated to be Reade’s late mother discusses an incident her daughter experienced with a current senator.
• April 26: Fox News reports that the episode in question was removed from Google Play, and the episodes re-numbered to go from 154 on Aug. 10 to 155 on Aug. 12.

The last point is particularly damning, although it’s unclear whether CNN or Google is accountable for it.

Twitter won’t let it trend

This story would seem to be hugely important, and the latest major development emerged just today. Yet Twitter isn’t listing it as a trending topic — not in the top 5, top 10, or even top 20.

There are many potential explanations, e.g. it could be a downstream consequence of some other factor, which itself could be innocent (e.g., Twitter users lean left) or nefarious (e.g., journalists aren’t reporting on it so people aren’t tweeting about it).

Unfortunately, data supports the least charitable explanation: there’s enough Twitter activity for the story to “trend,” but Twitter won’t allow it.

Here’s the proof. I recorded the top 20 Twitter topics at 4:00 pm, and then wrote an R program to count the number of tweets posted containing each term from 3:55–4:00 pm. I also included “Tara Reade.”

Based on Twitter activity, the story ought to be in the top 10, and certainly the top 20. But it isn’t and hasn’t been all day. By the way, I ran this analysis a few times today, and “Tara Reade” reached as high as #4 around 11 am. The above result isn’t cherry-picked.

Google Trends won’t let it trend

At 6:09 pm Eastern Time, Google Trends listed these topics as the top 14 for the day:

If you scroll up and actually search some of these topics along with “Tara Reade”, it’s easy to see they’re fudging the numbers. The Tara Reade story dwarfs many of these topics in search interest.

Unethical and sloppy work under the guise of “fairness”

Suppressing a story like this is extremely unethical for obvious reasons: it protects an alleged predator and serves to advance partisan interests.

It’s also naive and sloppy. Do Twitter execs think people will believe there’s more interest in a mascot being removed from a butter brand (Land O’Lakes) than allegations against Joe Biden? Do they not realize people can use Twitter’s own API to cross-check their trending list?

Google’s even sloppier — they have a curated “top 10” list, and all you have to do is scroll up on the same page to identify blatant omissions.

Ironically, big tech companies are constantly pushing the idea of “equitable” and “fair” algorithms. I think they use those nice-sounding terms to obfuscate things, so ultimately they can highlight whatever topics they prefer to highlight. When confronted with a question like “Shouldn’t the Tara Reade story be trending?” they can say “Hmm, well, apparently our fair algorithms…”

At a certain point, the algorithms aren’t even algorithms. There’s almost certainly a manual override, where someone has the opportunity to choose individual topics to suppress. I’d guess the general procedure is:

• Sort on an objective metric
• Remove topics that are inconvenient

One way or another, they’re finding a way to promote and suppress certain types of content. I think the Tara Reade omission is remarkably bold, and likely to lead to severe and irreparable backlash.