Decisions are power. The decision-maker has agency over the representation, framing, analysis and conclusions drawn from their research.
The process of decision making is abstracting – from broad, raw data, into a selection of options. The data is the input and the options are the output. I will use the example of choosing between juice and coffee to outline the process.
Traditionally a programmer would outline a specification for decision making in code, considering specific aspects of the input to create an output.
However, any level of complexity becomes quite complicated to manage…
A neural network is a form of A.I. that makes decisions. It combines data to create more nuanced information (located in the 1st layer), this process is repeated (within further layers) to create further abstractions until it is an output; juice or coffee.
Neural networks are based on the human brain. How they combine data is taught to them through trial and error. It is the influence the neurons give each point of data that makes it useful but also separates it from other networks which draw different conclusions from the same information. This is subjectivity. Selecting which aspects are important, which influence your decision, removes objectivity.
With this understanding of how decisions are made, we can see that any process – ordering, formatting, excluding information – all remove objectivity. While it means it’s not possible to create objective research, perhaps it is more valuable to document the role of our subjectivity as a tool, rather than disguise it as a pseudo-objectivity – acknowledging the power our decisions exert over our research.
