Is So-Called Contextual Advertising the Cure to Surveillance-Based “Behavioral” Advertising?

Contextual advertising might soon rival or even surpass behavioral advertising’s harms unless policy makers intervene, says Katharina Kopp.

Shutterstock

Contextual advertising is said to be privacy-safe because it eliminates the need for cookies, third-party trackers, and the processing of other personal data. Marketers and policy makers are placing much stock in the future of contextual advertising, viewing it as the solution to the privacy-invasive targeted advertising that heavily relies on personal data. 

However, the current state of contextual advertising does not look anything like our plain understanding of it in contrast to today’s dominant mode of behavioral advertising: placing ads next to preferred content, based on keyword inclusion or exclusion. Instead, industry practices are moving towards incorporating advanced AI analysis of content and its classification, user-level data, and insights into content preferences of online visitors, all while still referring to “contextual advertising.” It is crucial for policymakers to carefully examine this rapidly evolving space and establish a clear definition of what “contextual advertising” should entail. This will prevent the emergence of toxic practices and outcomes, similar to what we have witnessed with surveillance-based behavioral marketing, from becoming the new normal. 

Let’s recall the reasons for the strong opposition to surveillance-based marketing practices so we can avoid those harms regarding contextual advertising. Simply put, the two main reasons are privacy harms and harms from manipulation. Behavioral advertising is deeply invasive when it comes to privacy, as it involves tracking users online and creating individual profiles based on their behavior over time and across different platforms and across channels. These practices go beyond individual privacy violations and also harm groups of people, perpetuating or even exacerbating historical discrimination and social inequities. 

The second main reason why many oppose surveillance-based marketing practices is the manipulative nature of commercial messaging that aims to exploit users’ vulnerabilities. This becomes particularly concerning when vulnerable populations, like children, are targeted, as they may not have the ability to resist sophisticated influences on their decision-making. More generally, the behavioral advertising business heavily incentivizes companies to optimize their practices for monetizing attention and selling audiences to advertisers, leading to many associated harms. 

New and evolving practices in contextual advertising should raise questions for policy makers. They should consider whether the harms we sought to avoid with behavioral marketing may resurface in these new advertising practices as well. 

Today’s contextual advertising methods are taking advantage of the latest analytical technologies to interpret online content so that contextual ads will likely soon be able to manipulate us just as behavioral ads can. Artificial intelligence (AI), machine learning, natural language processing models for tone and sentiment analysis, computer vision, audio analysis, and more are being used to consider a multitude of factors and in this way “dramatically improve the effectiveness of contextual targeting.” Gumgum’s Verity, for example, “scans text, image, audio and video to derive human-like understandings.” Attention measures – the new performance metric that advertisers crave – indicate that contextual ads are more effective than non-contextual ads. Moments.AI, a “real-time contextual targeting solution” by the Verve Group, for example, allows brands to move away from clicks and to “optimize towards consumer attention instead,” for “privacy-first” advertising solutions.

Rather than analyzing one single URL or one article at a time, marketers can analyze a vast range o