Algorithmic Attention Rent

I was rather intrigued by this interesting piece of research by Tim O’Reilly and Mariana Mazzucato, amongst others.

The piece outlines a theory of algorithmic attention rents in digital platforms. As platforms and network FX grow, they become increasingly capable of extracting rents from a variety of actors in their ecosystem - users, suppliers and advertisers - through their algorithmic control over user attention.

Some of the core concepts and ideas discussed in the paper include:

• Algorithmic systems at dominant platforms like Google, Amazon, and Facebook control how user attention is allocated. This gives the platforms market power.

• User attention is a scarce resource. Platforms can extract "attention rents" by manipulating algorithmic rankings to direct user attention in ways that are more profitable for the platform but not ideal for users.

• Platforms can convert attention rents extracted from users into pecuniary rents extracted from advertisers and suppliers. For example, Amazon can use ads to take attention away from organic search results and charge suppliers for ad placements.

• Attention rents divert user clicks away from the most relevant organic results to more profitable but less useful sponsored results. This harms users, suppliers, and advertisers.

• Platforms should be required to disclose key metrics on how they allocate attention and monetize users. This data is needed to identify potential attention rent extraction.

• Metrics like the ratio of organic to paid clicks, ad loads, traffic referred to the platform's own sites vs third parties, and changes in ad pricing as growth slows can indicate attention rent extraction.

• Regulations could reserve screen space for organic results, require organic results to be ranked highest, or limit ad loads. But mandated disclosures of attention allocation metrics are likely more effective.

• As AI systems like chatbots replace search and recommendations, transparency is still critical to ensure fair attention allocation and prevent rent extraction by dominant platforms.

In summary, algorithmic attention rents allow platforms to misallocate user attention in ways that extract excessive value from users, suppliers, and advertisers. More disclosure of attention allocation metrics can help address this issue.