NYT is right to point out that Facebook’s recent announcement on US elections doesn’t go far enough, but their take missed some less obvious but very important problems.
Most obvious problem is that one week of political ads silence before the election is not nearly enough, particularly in a year when more people than usual are going to vote early because of COVID.
And it’s not even real silence, the word “new” in the press-release indicates that Facebook will continue to run the ads that will have been created before that week.
Also obviously, Facebook is not that accurate at labeling ads as political, and most of misinformation on their many platforms spreads through organic engagement that is often only ignited or amplified — but not powered exclusively — by inauthentic actors.
This part of the announcement is a token gesture that is only going to impact Zuckerberg’s faltering approval rating with his own employees, and not the integrity of the US elections in November.
Less obvious problem is with how instead of removing misinformation about voting methods (such as the Trump’s mail-in ballots post that they refused to take down in May) they will only add information labels (which proved next to useless with COVID misinfo) and will do that for any content that questions “lawful methods of voting.”
Voting machines are a lawful method of voting that is nevertheless extremely vulnerable to fraud. Are voting security researchers like Jennifer Cohn going to start getting misinfo strikes for reporting problems with voting machines?
Zuckerberg’s history of giving money to people who for decades enabled voting machine vendors to undermine security and integrity of US elections makes this a valid question to ask.
Steny Hoyer, co-author of the disastrous HAVA bill, is the rare kind of Democrat who regularly gets maxed out donations from Facebook PAC despite having less-than-bulletproof hold on his seat — a trait much more prevalent among Facebook PAC’s Republican recipients.
David Becker, who just received a $50 million grant directly from Zuckerberg, is a voting machine vulnerability denialist who gave cover to Georgia Secretary of State Brian Kemp’s effort to elect himself governor, and continues to work with the new Secretary of State Brad Raffensperger.
Last but not least, Facebook’s voter turnout tools are a Trojan horse that deserves a lot more scrutiny than it gets. Their impact can’t possibly be neutral between demographics that lean conservative or progressive. Conservatives being more likely to use Facebook and likely to spend more time on Facebook is already a source of bias. And the data on the voter turnout impressions categorization by age, gender, race, and districts doesn’t exist, probably even within Facebook.
Are underrepresented minorities more or less likely to see voter registration information? Are people in red districts of battleground states more or less likely to see it than people in blue districts? Facebook voter registration campaign can increase turnout in November by several percentage points, if it is biased in anyway that alone is going to be enough to decide the outcome of the presidential election and of many contested Senate elections.
The only way to prevent this from throwing the outcome of the 2020 US elections into question is the one thing Facebook is really bad at: transparency.