Who's Targeting Whom?
Happy Day Before Election Day! If you’re feeling frazzled by this country’s fractured realities and general inability to agree on anything, here’s something that might soothe a weary synapse or two: No matter how things play out on Tuesday and beyond, there’s already remarkable bipartisan consensus that “something has to be done about the technology sector.”
This general agreement was on display at the US Senate hearing I told you about last week. In broad strokes, both Republican and Democratic Senators expressed concern that America’s biggest tech companies are too big, too opaque about their practices, and woefully under-regulated. That’s not to say the hearing went well. Most of the coverage I read was justifiably negative, with many tech writers annoyed by the partisan theatrics and missed opportunities for digging into the complexities of Section 230, a controversial bit of federal code that made the modern internet and was supposedly the purpose of the hearing.
While I agree with much of the hearing criticism, I still found some value amid the Wrestlemania. Democratic Senator Amy Klobuchar of Minnesota, for example, spotlighted the hollowness of her Republican colleagues’ furious posturing by noting that it’s been four years since she introduced her bipartisan Honest Ads Act, yet the measure has gone nowhere in the Republican-controlled Senate. That law would, for the first time, bring meaningful federal regulation and transparency to online political ads, which, in addition to being one of the vehicles for Russian interference in the 2016 presidential election, continue to be a deep font of mystery and fury.
If Republicans want to actually regulate, rather than just fulminate about supposed online censorship while benefitting from online misinformation, they’ve had plenty of opportunity to intervene on the ad front and many others, Klobuchar was pointing out.
She didn’t stop there. While questioning Facebook CEO Mark Zuckerberg, Klobuchar noted his company at first offered “pushback” on the Honest Ads Act, later came to embrace it, and yet for all of Facebook’s record-level spending on lobbying in DC lately, the Honest Ads Act is still somehow stalled out.
“Have you done anything to get it passed?” Klobuchar asked Zuckerberg. “Because we’re at a roadblock on it.”
Unfortunately, Klobuchar stepped on Zuckerberg’s opportunity to answer by rolling straight into another question. So I put the question to a Facebook spokesperson after the hearing. No response.
When Zuckerberg tried to say Facebook already complies with the Honest Ads Act through its online Ad Library, Klobuchar responded that Facebook’s library doesn’t provide the level of detail on political ad targeting that her act requires.
It was good timing for this point to be made. Two weeks ago, the Ad Observer project, which is trying to sniff out the truth on political ad targeting, received a cease and desist letter from Facebook. According to Facebook, the project’s method of “scraping” ad targeting data violates Facebook’s terms of service. The Ad Observer project disagrees, and says targeting information is important for the public to have because of the unique qualities of online ads. Unlike ads in other mediums, the Ad Observer folks explain, “online ads are usually seen only by the audience the advertiser wants to target, and then they disappear.” (The Honest Ads Act describes this as their “ephemeral” aspect.)
Powered by the personal data American consumers are constantly handing over to Facebook and other digital services in daily interactions, the online advertising business is now able to target us with comic specificity. One could, for example, aim an ad campaign at just Volvo-driving, latte-purchasing, stay-at-home fathers between the ages of 32 and 52 in a particular zip code in Berkeley. Given the nature of such ads, whatever message those fabulous Berkeley men were receiving would, by definition, be kept away from the Facebook feeds of users not meeting the description. “This makes it difficult for the public to monitor [such ads] and hold advertisers, including political groups, accountable,” the Ad Observer project notes.
Zuckerberg told Klobuchar that “getting the right resolution on that is challenging without it becoming a privacy issue.” A Facebook spokesperson did not clarify what he meant by this, but it’s a defense Facebook has raised against required transparency for political ad targeting in Washington state, too.
Look for the Honest Ads Act, Facebook’s lingering concerns about it, and the broader issue of lifting the veil on ad targeting to become a focus as regulatory efforts ramp up next year.
Expect political ad bans to be discussed more next year, too. As I warned nearly a month ago, Facebook doesn’t have a great track record with these sorts of bans. That poor track record made me skeptical the company could cleanly pull off its new, nationwide ban on political ads ahead of this historically tense election and—yeah, it couldn’t.
The campaign of Democratic Presidential candidate Joe Biden sent out furious statements on Friday saying that “thousands” of their ads were stuck in some kind of un-viewable purgatory, even though they’d been approved to run before the new ban went into effect at 12:01 a.m. on October 27th. Biden’s camp also noted that Trump ads were still getting through, even when they were on the same subject as Biden’s stalled ads.
All of which recalls a key lesson from Facebook’s failed ad ban in Washington state: If Facebook is going to ban political ads, any ban must be airtight and evenly enforced, otherwise the company will be accused of effectively putting its thumb on the scale for one candidate over another.
By the way, what gives Facebook the right to both ban political ads and do a very imperfect job of enforcing such bans? Once again, it’s Section 230 (according to Facebook). Look for this particular slice of immunity to become an issue next year as well.
Finally, and still on the subject of future reforms to Section 230, Wild West subscriber Jon Lasser writes:
What if all algorithmic promotion of content (i.e., recommendations, “your friend liked,” “trending topics,” etc.) was considered speech *by the service/promotor* and therefore not subject to safe harbor [under Section 230]?
Having your content appear on your page, or even in chronological order on someone’s list, would be protected and seems to me at least substantially different than the outrage-generating algorithms that seem to be at the heart of the amplification problem.
Jon is onto something here. Shortly after I received his email, Roddy Lindsay, a former data scientist for Facebook, wrote an interesting piece for The Information arguing for one elegant reform to Section 230: make clear that the law doesn’t provide platforms immunity for “algorithmic amplification” of people’s speech (which, after all, is the platforms choosing which speech becomes loudest).
Lindsay writes:
A restriction on algorithmic content feeds could be a moment of modern human liberation. We’d finally have a policy ally to beat back our tendency to “doomscroll” or fall into Recommended For You ruts during our inevitable stretches of mental weakness.
Whether or not this is where Supreme Court Justice Clarence Thomas was headed with his unusual request earlier this month for an opportunity to rule on Section 230, the ideas that Justice Thomas outlined provided “an intriguing roadmap” for Lindsay’s own proposal. I’m willing to bet this concept of making platforms responsible for what they algorithmically amplify will come up at future Senate hearings that are, hopefully, more focused on the actual nuts-and-bolts of tech policy reform.
While we wait for Election Day and definitive vote counts, some stories worth your time:
• “Too worried” — Shannon McGregor and Daniel Kreiss argue that Americans have become overly concerned with political misinformation, to the point of using it as an excuse to blithely write off the sincerely held political opinions of others.
• “Definitely true… But…” — McCay Coppins of The Atlantic notes that “misinformation also hardens political identity—and is cynically exploited by powerful people who know better.”
• “Striking” — Julia Angwin on the differing rates that Biden and Trump are charged for similar Facebook ads.
• “Prebunk” — An interesting move by Twitter to highlight accurate information on the contested topic of voting security. (Sign of the Twitter times: It feels like this happened ages ago, but it only rolled out last week.)
• “Perception hack” — In case you want one more thing to worry about before the election.
Good luck to all of us tomorrow! Questions? Tips? Comments? wildwestnewsletter@gmail.com