Google wants changes to new state rules promoting political ad transparency.
Google, for example, doesn't want to be held responsible for knowing when an ad it sells is political. Justin Sullivan / Getty Images

Ahead of an important public hearing tomorrow, lobbyists for the nation's leading tech companies are urging Washington State officials to make major changes to powerful rules governing online political ads.

Among other things, the tech lobby's proposed changes would erase language in state code that requires companies like Google and Facebook to actually know when they've sold a political ad aiming to influence Washington's elections.

Here's a look at the wish list that tech lobbyists recently presented to the Washington State Public Disclosure Commission, which will be taking public comment on this issue beginning at 11 a.m. tomorrow in Olympia:

Tech Wish 1: No responsibility for actually knowing when an ad is political.

In a September 21 letter to the Public Disclosure Commission, Google argued that current state rules "place an enormous burden on commercial advertisers" by requiring them to know when an ad they sell meets the state's definition for "political advertising" or "electioneering communication."

The Internet Association—whose members include Facebook, Twitter, Google, and many other tech titans—made the same argument in its own September 21 letter, writing that "it is technologically impossible" for its members to know whether every single ad that runs on their platforms is political or not.

"Google serves tens of billions of ads every day," the company wrote in its letter, "including ads by state and local candidates who use self­-service Google Ads without ever interacting with anyone at Google."

In June, Google announced it would no longer sell political ads in Washington State because it's not prepared to meet the requirements of current state transparency rules. However, in a development that seems to confirm Google is not flawlessly tracking whether every ad it sells is political, the company nevertheless sold more than $6,000 in political ads to Washington State candidates in recent months.

In its letter to the Public Disclosure Commission, Google goes on to brag about the company's technological prowess, noting that "Google's machine learning algorithms can analyze up to 70 million signals within 100 milliseconds, quickly interpreting all possible data combinations, and then rapidly evaluate vast sets of data signals to proactively set the optimal bid and ad placement, often in real time, for each advertiser."

Which raises a question:

Why is it that a company capable of such feats cannot determine when it has sold a political ad that should, in the interests of informed democracy and accountability, be given a certain amount of human scrutiny before it is published and a certain amount of transparency after it is published?

Google's letter does not firmly address this question, though both Google and the Internet Association took pains to present themselves in their letters as champions of election transparency.

"The people of Washington state deserve more information about the political communications they're receiving online," Google wrote.


Tech Wish 2: Consider the market desires of tech titans and, if you don't, worry about a loss of "free speech."

Google, in its letter, claims that current state rules "risk driving commercial advertisers out of the market for political communications in Washington." The Internet Association, in its own letter, made a similar argument.

Of course, as the letters also acknowledge, it will be up to the Public Disclosure Commission to determine whether its most important goal is helping companies like Google and Facebook sell ads with as little friction as possible.

Created more than four decades ago by a vote of 72 percent of Washington State citizens, the commission's primary aim is to bring transparency to state and local elections as a way of strengthening democracy and increasing public knowledge about the money behind political messaging.

The Internet Association also told the PDC that leaving the current transparency rules unchanged, and potentially leading tech companies to abandon the Washington State political ad market, "would have a stifling effect on free speech."

This argument leaves out a key fact: The ads at issue are paid speech.

The paid speech in question—again, online political ads—is mediated and controlled by private companies that are under no obligation to offer their advertising clients anything close to what Americans typically think of as "free speech."

Tech giants are not the public square, even if they've been allowed to become so large and omnipresent that they sometimes seem as if they are.

They are more accurately described as massive digital platforms that privilege paid speech in order to make a profit (and also privilege emotional appeals and incendiary rhetoric through their focus on getting users to like, retweet, share, and otherwise promote their content).

Companies like Facebook and Google can choose where they want to offer customers opportunities for paid speech.

But if they choose not to offer opportunities for paid speech in Washington, that doesn't at all mean that free speech is dead in Washington—or even stifled.


Tech Wish 3: Make ad purchasers responsible for accurate disclosure, not tech giants.

Both the Internet Association and Google argued to the Public Disclosure Commission that the burden for identifying political ads should be taken off the shoulders of the tech titans.

Instead, they said, it should be placed squarely and solely on the shoulders of the individuals purchasing political ads—with tech companies granted a "safe harbor" to merely report whatever an ad purchaser might tell them about whether or not a particular ad is political.

Support The Stranger

The reason tech companies want this "safe harbor" is their fear that under Washington law, they have to be right 100 percent of the time about 100 percent of their disclosures for 100 percent of the political ads they sell. (Or face, say, a lawsuit from Attorney General Bob Ferguson.)

But under the tech companies' proposed remedy, if an ad purchaser lies or otherwise misrepresents their political ads to Facebook or Google—not a far-fetched or hypothetical scenario given the 2016 presidential election—then we'd quickly have a "garbage in, garbage out" problem in Washington State.

The shady advertiser would lie to the tech company, and the tech company, safe in its new "safe harbor," would still be in perfect compliance with state rules if it just went ahead and parroted that shady advertiser's lie in its own public disclosures.

To put it another way, this would represent a perfect closed circle of bullshit.

For a clear example of just such a circle in action, see this year's campaign to repeal Seattle's "Amazon Tax" by local referendum.

The "Amazon Tax" repeal campaign reported to the PDC that it had spent $1,250 on Facebook ads and $1,250 on Twitter ads.

In its own public disclosures, Facebook didn't offer any information to contradict this claim.

But the claim was false—something I was only able to discover because Twitter, unlike Facebook, complied with current state rules and independently verified that $0 had in fact been spent on Twitter ads by the "Amazon Tax" repeal campaign. (Rather than the reported $1,250.)

That, in turn, led to claims that the "Amazon Tax" repeal campaign had been scammed by a sub-vendor. It also led to a new, revised report to the PDC about the campaign's Facebook ad spending—which turned out to be just $374.81 on Facebook ads, not the reported $1,250.

If tech companies are free to simply parrot whatever their clients say about their ad purchases, this kind of stuff doesn't get exposed—even when it happens on a much larger scale.


Tech Wish 4: Don't make digital platforms disclose too much targeting information for online political ads because that'll give away the game.

Current state rules say companies like Facebook and Google need to disclose exactly how much has been spent on a political ad, who paid for it, and exactly how many impressions an ad generated.

Such companies also need to disclose an "approximate description of the geographic locations and audience targeted" by the online ad purchaser.

"Notably," Google writes, this last requirement "extends only to digital ads, not to broadcast or print ads."

True enough, but what Google neglects to mention is that digital ads purchased through Google (and Facebook) are unique in that they can precisely target small, specific groups of potential voters in ways that old media ads cannot.

This capability was not only demonstrated by the Russians who interfered in the 2016 presidential election—skillfully pitting one specific demographic group against another using tailored, targeted Facebook ads—but also by the 2017 campaign of Seattle city attorney candidate Scott Lindsay, who used targeted Facebook and Google ads to push a misleading claim about property crime on specific Seattleites who lived in specific Seattle neighborhoods.

I only found out about Lindsay's ad targeting because I asked him on a hunch and he admitted it to me.

While defending his tactics, Lindsay also admitted that "very clearly abuse is possible" using targeted political ads.

Robust transparency around online political ad targeting is one obvious way of preventing such democracy-warping abuse.

But Google, in its letter, counters that "depending on the breadth of this requirement, such disclosures will force commercial advertisers to reveal confidential, strategic information about campaigns' operations." Political campaigns, Google warns, might prefer to just stop buying online ads altogether "rather than face public disclosure of their confidential, campaign strategies."

So, Google argues, "an appropriate balance of sponsors’ and the public’s interests is to require commercial advertisers to disclose the gender, age range, and geographic location of the audience the advertiser seeks to reach. This provides sufficient information for citizens and users to understand why an advertisement is reaching them without giving away a candidate’s 'playbook.'"

The devil is in the details here.

For example, the "geographic location" Facebook eventually decided to disclose for Lindsay's ads was merely "Washington"—for a Seattle-specific race in which Lindsay's ads had, in fact, been targeted at specific Seattle neighborhoods and specific Seattle voters. (Google, for its part, didn't disclose any targeting information at all for Lindsay's ads.)

The major danger of targeted online ads derives from the same source as their power, which is that such ads are effectively "dark" to everyone who's not targeted.

As targeting abilities grow even more powerful and precise—further increasing how "dark" such ads can get—the daylight antidote may need to become more powerful, too, involving significantly more categories than gender, age range, and geographic location. Given how powerful ad targeting capabilities are already, the antidote may need to encompass more that just those three categories right now.

Perhaps the Public Disclosure Commission, as it's exploring this particular question, should determine exactly how many ways there are for a political advertiser to target an ad beyond just gender, age range, and geographic location.

But at a bare minimum, it seems obvious that this kind of disclosure should definitely include specific data on how many people were actually reached by any online political ad.

Google, however, argues against this. "It is impracticable to provide a real­-time accounting of the precise number of impressions generated by a communication," the company wrote.

Notably, "impracticable" has a different definition than "impossible."

Also notable: Elsewhere in its letter, Google explains that when it sells ads, "the advertiser is only charged once the ad has been viewed or clicked upon."

Which raises the question: If Google can track an ad's reach precisely enough to charge clients by the view or click, why can't it total up views and clicks for prompt disclosure purposes?


* * *

There are many other issues raised by the Google and Internet Association letters, which, if you're a political or tech nerd, are well worth reading in full.

Few of these issues will be easy to resolve to universal satisfaction, and the PDC—as the Internet Association notes in its letter—will have to wrestle with all of this in the context of the overworked agency's "limited funding."

But Washington has long been a leader on election transparency, and pretty much everyone involved seems to agree that it's important to democracy that the current rules—if they do need revising—be revised carefully and with the public good foremost in mind.

Kathy Sakahara, a lead election transparency advocate for the League of Women Voters of Washington, plans to testify at the hearing on Thursday morning.

"The entire point of these rules is transparency," Sakahara said. "That's why the League was instrumental in founding the PDC and why we continue to support the PDC's efforts. I look forward to making the case for continued transparency and fighting for the right of the public to know who is paying for campaign ads. We will continue to fight for that, at this hearing and beyond."