Ad Algorithms Sell Cheap Impressions Regardless of Real Attention
Algorithms optimized for advertising revenue have made impressions the main unit of value, even though genuine audience attention is often absent. Now AI is intensifying the problem by filling feeds with automatically generated content.
Algorithms built around advertising revenue have turned audience attention into a cheap commodity, with showing an ad becoming more important than whether anyone noticed it. This is the central thesis of the material on how advertising machines have undermined the attention economy. According to the notes, this model is now deepening further through artificial intelligence, which can create an unlimited amount of content and automate the creation, targeting, and purchasing of advertising.
The story began with the shift of advertising budgets from television to digital platforms. The television advertising market was in trouble: cable subscriptions were declining, ratings were falling, and advertisers began moving billions of dollars to where they were promised a cheaper and larger audience. In 2016, a BuzzFeed video about rubber bands and watermelons gathered more than 3 million views. At the same time, these views could include short autoplays, so the real attention of viewers remained unknown.
The reason for this uncertainty lies in how the industry measured effectiveness. The industry counted a view lasting three seconds as an impression, and the impression was the basic unit of cost. Algorithms were optimized for advertising revenue rather than genuine attention, so impressions were counted even when nobody cared about the content. Over time, these systems expanded their claims and began taking credit for purchases people would have made anyway, without the ads.
The result was a sharp growth of advertising inventory. The supply of cheap, low-quality impressions flooded the market, and digital advertising ended up in the same price category as television advertising during peak viewing hours, even though their levels of attention differed substantially. Television networks responded to the pressure by increasing the number of advertising minutes. This worsened the viewing experience and pushed audiences toward streaming services, creating a closed cycle.
Attempts to change the situation on television were limited. Fox Networks Group, headed by Jonathan, tested six-second spots and premium advertising placements. Advertisers, however, did not want to pay more for shorter and more premium formats. Brands preferred volume over the placement and scarcity of slots, so the incentive to produce quality advertising remained weak.
Now this logic is being carried into the era of artificial intelligence. It generates unlimited content and fills feeds with AI-created videos, while platforms automate the creation of advertising, targeting, and purchasing. According to the notes, ChatGPT is reportedly approaching an annual advertising revenue of 1 billion dollars, and more than 100,000 advertisers are allegedly already launching campaigns on the platform. The authors warn that this could lead to even cheaper and less meaningful impressions, while values that are hard to measure will be left without attention.
At the same time, the authors of the material express the view that AI can be built to help people manage attention: filter information, manage their time, and support better decisions. This idea is at the center of the proposed reset, but the specific mechanisms for implementing it are not described in the notes.
Source: fastcompany.com
12
