Diversity of Thought & Education Matter

The vast majority of data scientists and software engineers understandably come from STEM backgrounds. How are they supposed to think rigorously about model positionality (social situatedness of sociotechnical systems) and the ethics of the sociotechnicsl systems they build when they have little training in such modes of thought?

It’s not a shortcoming on their part not to think about ALL the ways the system they’re building could cause harm, especially given how subtle and complex such matters can be. It’s the job of management to bring a diverse skill set, which includes diversity of thought and education, to the system building process.

Bringing in professionals with expertise in how to mitigate risk and find hidden, potential harms demonstrates a good faith effort to do no harm and build systems responsibly.

And, critically, it’s easier (and usually cheaper) to build with safety and ethics from the beginning than to ask an ethicist to bring a mop and clean up the system’s issues after the fact.

Required audits are coming—and are actually here in a few contexts. It’s just a matter time before audits grow exponentially. Build knowing your system will likely be audited in the future by using a team with diversity of thought and education.