I see my three fields of Visualizing Computation, Critical Computation, and Reflexive Computation, as three components of a methodology.
I imagine developing a dissertation project that puts into practice the ideas I have been exploring in my three fields, as a design research project.
My main idea is to work with a single algorithm that is used in various contexts (terrorism, spam, welfare, shopping), from the controversial to the benign. Find a foundational algorithm, visualize its logic, then show how it works in a benign context, and also an ethically controversial context. How is it different, where is the ethical position in the computation? Visualize the difference.
How to visualize an ethical position in an algorithm? I haven’t done this yet, other than mistakes.
Working with the Center for Artificial Intelligence in Society, at USC
A lab at USC that I have been working with, led by Milind Tambe.
I am unsure how much my disseration project should be defined by the lab.
They have a suitable project that addresses ethics in AI. Algorithms that use bayesian probability in network theory, applied to vulnerable groups in Los Angeles.
Moving from a discussion of ethics to a discussion of human rights in algorithms. Philip Alston, NYU School of Law
What would a propositional, speculative AI look like, for affirmative action for example? Bias/mistakes presumes neutrality is possible. How to visualizing the asymmetry of power vs visualizing bias?
How does the culture feel it? How can I bring back cinematic, sensual imagery into my work, in combination with the graphical and analytical? Should I?
Should I work on deep learning?
Visualizing how emergence is a political issue?
Kevin De Liban, Legal Aid of Arkansas, lawyer litigating machine learning algorithms used to determine access to medicaid.