Morality in the Face of Algorithms
Like a rippling pond, more advertisers are pulling more money from their YouTube ad spend over an ever wider net of issues.
Since mid-March, a series of large advertisers on Google’s video site YouTube have announced their intention to reduce or remove their ads after it was revealed they were showing next to and supporting extremist content.
The rot started with a report in the UK’s The Times on March 17, and spread from the companies named there to others in the UK and Europe. Shortly afterward, companies in the US including AT&T, Verizon, and GlaxoSmithKline “suspended advertising spending on Google except for search”. The global moves have also spread to Australian subsidiaries, according to the Australian Financial Review.
This issue has come about because Google uses a method of placing ads alongside user-generated content on YouTube and other sites called programmatic advertising, which allows the process to take place efficiently, at scale, and cheaply. The brands now boycotting Google’s ad products, their spokespeople, and marketers, are making entirely rational and moral decisions in response to what are effectively amoral algorithmic actions.
That is not to say that the algorithms (or any technology) is value-free, just that algorithms themselves are not making moral judgments when performing their function. There’s a dangerous recipe involved in equating the two. Jean Baudrillard, in Simulacra and Simulation, writes “the meticulous operation of technology serves as a model for the meticulous operation of the social.” Allowing ourselves to be convinced of the infallibility of technology convinces us also that humanity can be controlled and made infallible.
Clearly, much of the content cited as reason for ad-removal decisions is widely objectionable. However, the rapid response of advertisers in reducing their spend suggests a more deep-seated concern. It suggests that these issues have validated what many advertisers and companies have long suspected of programmatic advertising — although it is the way to do business online, the algorithm can’t be trusted. These concerns stem in part from the not-infrequent criticisms of the way online views are counted. (YouTuber Hank Green’s piece ‘Theft, Lies, and Facebook video’ is an excellent primer on this issue.) They are reflected also in claims that “60% of programmatic spend is wasted”.
The companies concerned have also shown themselves to be open to coercion in decisions about buying and placing ads on particular sites or through particular ad networks. Given popular and government concern about the terrorism and anti-social content which sparked this particular incident, this is not surprising. However, the pressure can extend to other interest groups from less-widely supported movements.
How can Google fix these issues?
Many of the companies quoted in the stories linked above say they are working with Google to fix the issue. This week, tech researcher danah boyd argued that we can’t necessarily fix big-platform issues like ‘fake news’ (ugh) with “algorithms or editors — because the problem is….us”.
A blog post from Google’s Chief Business Officer Philipp Schindler repeatedly highlights the importance of “values” in fixing the problem. But the solution is much more prosaic. They’ll implement further human controls on deciding which content is able to be monetised (that is, available for ad-placement) and give advertisers more options to reject particular channels or keywords from their buying schedule.
Finally, Google promises to they’ll be “hiring significant numbers of people and developing new tools powered by our latest advancements in AI and machine learning to increase our capacity to review questionable content for advertising.” The solution, then, is a mix of human and more human-like machine decision making. It remains to be seen whether the AI aligns with advertiser values.
Other relevant articles:
