Game or measurement? Algorithmic transparency and the Klout score

Devin Gaffney

29 Sep 2012 — 1 min read

Cornelius Puschmann and I wrote up an interesting and short paper about serious complaints with Klout, a now less important, but at the time very important website dedicated to scoring the influence of individuals. In short, we argued that influence is inherently contextual, that a commercial platform necessarily prioritizes monetization over any legitimate measure, that an opaque algorithm is problematic, and that users need agency in defining their influence. You can read a short version of the paper below and view my presentation below along with the abstract.

YouTube

Abstract

Klout is an Internet service that claims to measure an
individual′s influence by aggregating information from a variety of social media platforms. We briefly examine Klout′s approach to measuring influence and argue that the Klout score′s lack of algorithmic transparency undermines its status as a trustworthy metric. We further argue that Klout and similar services “gamify” the notion of influence in ways that encourage competitive behavior among users in ways which are detrimental to the quality of measurement in a scientific sense, instead encouraging a gamified notion of influence. We conclude that only an approach to measurement that is based on transparency has a real chance of gaining trust, and that only a widely trusted metric may serve as a reliable indicator of influence. Beyond the case of Klout, context, theoretical soundness, algorithmic transparency, and user agency are general issues that must be dealt with for any entity attempting to measure influence in online social networks

influence12_submission_23

influence12_submission_23.pdf

37 KB

The Future is Algorithmic Feeds on Bluesky

If the original sin of Web 1.0 was the pop-up ad, the original sin of web 2.0 was the move to algorithmic feeds. Opaque optimization strategies aimed at maximizing private revenue for the sake of what was otherwise externally billed as public goods became increasingly toxic, spawning discourse

How Much Data is Enough for Finetuning an LLM?

There's no shortage of analogies for explaining what an LLM is capable of - one of the best, though, is from this New Yorker article proclaiming it as a "blurry JPEG of the web". This metaphor is particularly useful for capturing many of the technical aspects

Using Synthetic Data Generators to Measure LSTM Lift

Long short-term memory models (LSTMs) are a family of neural networks that are predominantly used to predict the next value given a historical chain of previous values. These can be numerical predictions (i.e. where is the stock price going based on historical stock data) or categorical predictions (i.e.

Some supervision required: LLMs at scale in practice

Recently, I gave a talk at the PIE/Autodesk space to help contextualize some thoughts that have been percolating with regards to the nascent introduction of API-based, widely available LLMs like ChatGPT. In the hype cycle, I've observed some pretty broad claims about what's happening under

Read more

The Future is Algorithmic Feeds on Bluesky

How Much Data is Enough for Finetuning an LLM?

Using Synthetic Data Generators to Measure LSTM Lift

Some supervision required: LLMs at scale in practice