Embed Notice

@ned @lore I was faced with questions of “ethical data analysis” some years ago, and my view was the same: it’s not the question, and it’s not the answer, it’s the what you do with the information.

I’ve since shifted a bit, I now think you can ask if a piece of work is ethical to engage with based on likely outcomes and the decision making or functions it enables.

A radical example: Bump Stocks were created to increase semi automatic rifle fire rates to automatic fire rates, there’s no practical domestic use beyond entertainment value. Was the risk position of these products in the wild assessed before building and productionisation? It’s the US, so what knows. After Los Vegas and 860 odd direct and indirect casualties the sale of these products was suspended, for a month, until they were used again before permanent sales suspension. There’s no governmental ban, the manufacturer decided it was unethical to continue sales, and even defended their patent to prevent another manufacturer picking it up.

Should Slide Fire have predicted the risk the the community? Yes. Should they have done something to mitigate the risk (including filing the patents and never building bump stocks)? Well, yes. They made the right decision, but only after ~1 thousand people were injured and scores killed across two incidents.