“Where do I go to get my reputation back, even though nobody knows I lost it?”
That seems paradoxical, since your reputation is what people generally think of you — but in digital publishing what really counts is what computers think of you. Earlier this week, secret algorithms besmirched my digital reputation, in ways that could negatively impact my paycheck, and my employer’s profits.
And it all happened behind the scenes, without any of my readers knowing. To you, I’m still the same old VodkaPundit with the same old reputation. For whatever that’s worth…
So what happened?
On Tuesday I wrote a generally well-received column, but I’m not going to link directly to it, nor am I going to tell you what it was about. The reason why will be clear in a moment.
ASIDE: You can find it on my columnist page here, the only entry dated June 18.
I will tell you this much: It was a column looking at the pros and cons of artificial intelligence enabling the manufacture and use of ever-more realistic robots dedicated to… knitting. (Not really knitting, but I have to be ambiguous for a short while longer.) These AI-enabled “knittingbots” could have all kinds of benefits for lonely people who have to do too much of their own knitting, or for people who previously paid money to professional knitters who weren’t always willing knitters, and others. But, as I discussed in my column, knittingbots could also exacerbate ongoing negative trends concerning the, ah, fecundity of yarn-makers.
If you catch my drift.
Shortly after publishing my knittingbots column, I gave our managing editor, Paula Bolyard, a heads-up in case she wanted to feature it on the PJMedia home page. Paula got back to me at once, saying, “We’ve got a bit of a problem.”
The problem is that if you write about knittingbots, even in a completely non-prurient fashion, then stupid advertising algorithms destroy your digital reputation.
NOTE: If you already understand how web ads work, you can safely skip this next short paragraph.
What happens is that one company’s algorithm scans your article for content. It then shares those results with various other algos looking for smart places to place their company’s ads. If your content matches up with an advertiser’s target audience, the algos enter a brief (milliseconds) negotiation over the price, ads are placed on your column, and you make a little advertising money.
But algos are dumb. The can scan my perfectly innocuous column about knittingbots, and even though the content is actually rather dry social issues, decide that what I’m actually peddling is smut. PJM decided to demonetize the knittingbot article to avoid running afoul of the algorithms, which could affect the entire site. The net result is that my knittingbot column is generating lots of traffic, but it’s actually costing my company money to host it.
You know I don’t write that kind of stuff, and you know that PJMedia doesn’t publish that kind of stuff. And yet we, and other companies just like us, lose money every single day because of this digital defamation.
We’ve also been forced to a place where we can’t discuss important social, legal, and technological issues without having to use ridiculous euphemisms like “knittingbots.” Not, that is, if we want to stay in business.
It’s fraud, too. Algos are representing my work and PJM’s publication as something they’re not, in ways that hurt us financially, and presumably to the benefit of just-as-clean sites that haven’t been red-flagged. While I can’t prove anything, due to Google’s algo being both opaque (no one but Google knows what’s in it) and invisible (you can’t see it working), I somehow doubt that the major, left-wing publications get the same red-flag treatment. The New York Times has published a number of pieces about knittingbots dating back to 2007, but it’s a safe bet that Google hasn’t tagged the entire nytimes.com domain with the ruinous red flag.
What to do? Proving defamation is nearly impossible in this country, especially the digital variety. Antitrust is fraught with risk, too. A few weeks ago I wrote a piece (“Hotel Googlefornia“) concluding “Break. Them. Up.,” but even so I continue to go back and forth on the issue. As I snarked on Facebook the other day:
What You Want from Antitrust:
• Restore competition to the tech sector, particularly in internet communications
• Resulting in increased privacy and faster innovation
What You’ll Get:
• Republicans will realize they don’t know what the hell they’re doing, give up, and let the Democrats call the shots
• Democrats will then use it as an opportunity to shake down the FAANG corporations for big money, with the implicit understanding that they won’t shear anyone too close
• Status quo ante antitrust, more or less
And Republicans seem determined to prove my cynicism correct. There’s a push on to kill Section 230 of the 1996 CDA law, but that could prove to be a mixed blessing — at the very best. Elizabeth Nolan Brown warns that you can “Expect More Conservative Purges on Social Media If Republicans Target Section 230,” and David French, whom I disagree with on a number of issues, nevertheless appears to be on the mark here:
It’s often the case in Washington that the title of a bill communicates the exact opposite of its content or effect. Think, for example of the Affordable Care Act — a title that seemed almost laughable in the face of skyrocketing insurance premiums. Now we have the Republican version of a deceptively named bill, Missouri senator Josh Hawley’s Ending Support for Internet Censorship Act.
In reality, it’s a bill that would inject the federal government directly into the private social-media business and grant it enormous power over social-media content. It would enable public censorship in the name of limiting private control.
What then, is to be done? Take another quixotic defamation case to court? Try and find the financing — and enough sympathetic tech talent — to break Google’s stranglehold on internet advertising? Risk breaking the internet by getting big, fat government to sit on its back? Or just continue to suffer the slings and arrows of outrageous algos?
These are tough questions, and there might not be any truly satisfactory answers. While I stew on them, I welcome your input in the comments.
Just keep it clean, OK?
This article was updated to include a better explanation of the demonetization of the knittingbot article.