Ahmad, Sens, Eesley & Brynjolfsson have just published a study in Nature that shows how ill-informed most corporate advertising execs still seem to be about where their online ads are appearing, an ignorance which also appears to be funding a variety of misinformation-oriented websites.
I noted in Chapter 10 that Google has made it increasingly challenging for advertisers to apply traditional search engine optimization techniques to their ad buys, so this potential side-effect of that is not at all surprising to me. And here’s a new article in The New Yorker showing yet another side-effect of Google’s secretive actions: this one impacting content creators. I can only hope Google search liaison Danny Sullivan reads and responds to these articles, since I’m sure that his peers at Meta, etc., whoever they are these days, won’t.
The Nature article’s “supply-side” approach to “defunding” misinformation is also a novel addition to the field, since most misinformation and fake news researchers have focused on users. One quibble I have with this study is that they studied both consumers and advertising executives as separate parts of this project, but apparently they only solicited and analyzed political preferences from the consumers, finding that the politically “liberal” users tended to be more averse to the misinformation websites. It seems to me that the advertising executives should have been asked about their political preferences too, since this tendency may have shown up there as well. Such a finding could potentially change the study’s title from “Companies inadvertently fund online misinformation despite consumer backlash” to “Some companies inadvertently fund online misinformation despite consumer backlash while others do it on purpose.”
Anyway. From their conclusions section:
“On the basis of our results, we identify two interventions that could reduce the financing of online misinformation. First, digital advertising platforms that run automated auctions could enable advertisers to more easily access data on whether their advertisements appear on misinformation outlets. This would enable advertisers to make advertising placement decisions consistent with their preferences rather than inadvertently financing misinformation. Second, while it is currently possible for consumers to find out about companies financing misinformation through media reports, digital platforms could improve transparency for consumers about which companies advertise on misinformation outlets. Platforms could provide such information to consumers when they are viewing an advertisement using simple information labels (as in our ‘company only’ information treatment) similar to the ‘sponsored by’ and ‘paid for by’ labels that are presently common on various digital media platforms. Similarly, rank-based information provided in our company-ranking information treatment (T4) could be provided as a ranking of companies in order of intensity of appearing on misinformation websites where customers are selecting products from a menu of choices while shopping. Platforms have provided similar contextual information about companies in other settings—for example, Google Flights displays carbon emissions data alongside flight prices when people select a flight to purchase among several options Enabling consumers to view such information at the point of purchase could provide a stronger incentive for companies to steer their advertisements away from such outlets, especially since the effect of negative information can persist for several months. Overall, these interventions could decrease the inadvertent advertising revenue going towards misinformation outlets, which could eventually lead to such sites ceasing to operate, as observed anecdotally in prior work.”
One can only hope!
Keywords In and Out of Context 