4 ms·

But you need somehow oversight right? (specially for something as critical as social networks used by billions) I'm a believer in some sort of cross-validation

But you need somehow oversight right? (specially for something as critical as social networks used by billions) I'm a believer in some sort of cross-validation and quantification. The ethical impact of a technology should be quantified, and addressed if significantly negative (or promoted/incentivized if positive!) -- you can quantify number of users impacted, give various arguments, soft measures and estimates. Then, you can have other teams validate those estimates. After all said and done, you can again evaluate an intervention (on both ethical terms, and on the productivity and profitability of your company), and see if it was good or not. If a team is repeatedly giving bad advice, I think you can then address that (retraining, lay off, etc.?). Ethics is not impossible to do :) I believe a certain generalization of this is needed for the entire society[1], to externally address good/bad technologies. A good example I cite all the time is Open source contributions. We still haven't found a good mechanism to pay OSS, despite the immense value it brings to society (openness itself has a value that direct sale couldn't capture). I'm a big believer in distributed evaluation and cross-evaluation: if we had distributed entities to evaluate impact of various ventures (for efficiency, properly categorized and subdivided into science, technology, social costs/benefits, environmental costs/benefits, etc.) we could apply the same logic to make society more adaptive, distributed, and less focused only on the single goal of profit. (I plan to elaborate this further, I think we are going to need to "patch" our current system in some sensible and careful ways to get a really good 3rd millennium!) [1] Previously sketched this here: https://news.ycombinator.com/item?id=28833230 https://news.ycombinator.com/item?id=28833230

duxup 4y ago
I think the problem is that the way these groups work ends up being something other than intelligent oversight. It's a human problem. They attract or breed people who see it as a great / easy job simply to provide oversight but have no investment / zero consequences for what happens. They can give tasks, eat up time, it costs them nothing, but costs everyone else tons of time. They are also free of any consequences ... how do you know such a group did any good? It's entirely for them to define. "These products were ethical, and we made them that way." Easy to say. IMO these kinds of questions are up to the folks in charge of the products, beyond just a random group deciding "is this ethical" as they actually also have to understand and know them. If the folks making the product can't handle it, that's the problem.
- AstralStorm 4y ago
  Perhaps if being ethical costs a lot of time, your business is unethical by design. Or worse, you're cutting corners that should not be have been cut. Now in terms of this particular abuse prevention case, the problem requires a serious social and technical solution that would be expensive for Twitter to implement, and would harm their revenue to boot...

duxup 4y ago

I think the problem is that the way these groups work ends up being something other than intelligent oversight. It's a human problem. They attract or breed people who see it as a great / easy job simply to provide oversight but have no investment / zero consequences for what happens. They can give tasks, eat up time, it costs them nothing, but costs everyone else tons of time. They are also free of any consequences ... how do you know such a group did any good? It's entirely for them to define. "These products were ethical, and we made them that way." Easy to say. IMO these kinds of questions are up to the folks in charge of the products, beyond just a random group deciding "is this ethical" as they actually also have to understand and know them. If the folks making the product can't handle it, that's the problem.

AstralStorm 4y ago
Perhaps if being ethical costs a lot of time, your business is unethical by design. Or worse, you're cutting corners that should not be have been cut. Now in terms of this particular abuse prevention case, the problem requires a serious social and technical solution that would be expensive for Twitter to implement, and would harm their revenue to boot...