Ancestry tracking services often deliver results in racialized “percentages”; algorithms render medical decisions without necessarily revealing the reasoning for their conclusions; built-into-the-machine algorithmic assortments tag disease and contagion with socially stereotyped categories such as race, religion, and nationality. The question for this session is how we might unpack data about public health in ways that do not re-inscribe the baggage of past historical division.

From Louis Agassiz to Nicholas Wade, from polygenism to the culture wars, we live in a world both segregated and bewildered by problems of proof. Whose truth reigns in a competitive universe that pits a.) reproducibility of results against b.) procedural rules prohibiting double jeopardy against c.) data privacy against d.) normativity and “common sense” against e.) sanctified canons, ritual practices, non-measurable belief systems, and the blinding complexities of what we mean by prejudice.

Here we will consider the degree to which this convening implicates translational issues. We will address linguistic/symbolic/representational confrontations between computational narratives and common parlance; between data complexities buried in quantitative reductions and peer-reviewed studies re-presented as popularized sound- bites; between knowledge production as “free” and “R&D” as productivity-oriented salesmanship; between private profit-motivated outcome and outright state-sponsored propaganda.

No abstract available.

The pervasive use of algorithmic risk assessment complicates or undermines many bedrock democratic values, institutional standards of accountability, and constitutional principles. Who –or what-- determines our fitness for inclusion within newly-constituted technological communities? What has primacy in what contexts—our personhood as defined by juridical or political consensus? Our status as corporate “stakeholders” in business enterprise? Our “risk factors” as governed by probabilistic modeling? The answers to these questions affect research standards and IRBs, notions of privacy as well as public transparency, fiduciary relationships with professionals like lawyers and doctors, the right to be free from unwarranted search and seizure, the right not to testify against oneself, freedoms of assembly and association, and conventions against non- consensual human experimentation. Ungoverned, data-driven assortments are creating new forms of stigma, disparate impact and group discrimination. We will explore what process is due and the merits of “explainability/interpretability/transparency” as a response.

No abstract available.

Can thicker descriptive and evaluative concepts in data analysis render richer results – or "fairness through awareness?" What correctives, cautions or constraints might there be for what happens when poor data informs algorithms that are planted like organizational seeds in the real world? This last panel will consider the social problems of profiling, stereotyping and civic notions of autonomy.