When models are everywhere
Last updated: 2022-11-29 Tuesday
When Models Are Everywhere
Hugo Bown-Anderson and Mike Loukides describe the everyday models that you interact with.
[Models are everywhere] They differ fundamentally from each other along dimensions such as alignment of incentives between stakeholders, “creep factor”, and the nature of how their feedback loops operate.
To understand the menagerie of models that are fundamentally altering our individual and shared realities, we need to build a typology, a classification of their effects and impacts. This typology is based on concepts such as the nature of different feedback loops in currently deployed algorithms, and how incentives can be aligned and misaligned between various stakeholders. Let’s start by looking at how models impact us.
Youtube says they show you what you want. But that is not true.
YouTube’s algorithm was measuring what kept viewers there the longest, not what they wanted to see, and feeding them more of the same.
authors propose to classify models with:
- nature of feedbackloops
- the alignment of incentives between stakeholders (model builders, users, and advertisers)
- “creep” factor
- “hackability” factor
- how networked the model itself is (is it constantly online and re-trained?)
what do I think about it
I wish they actually wrote out how they used these dimensions, I guess this is a first try.