Slashdot Mirror
Giving Algorithms a Sense of Uncertainty Could Make Them More Ethical (technologyreview.com)
An anonymous reader quotes a report from MIT Technology Review: Algorithms are increasingly being used to make ethical decisions. They are built to pursue a single mathematical goal, such as maximizing the number of soldiers' lives saved or minimizing the number of civilian deaths. When you start dealing with multiple, often competing, objectives or try to account for intangibles like "freedom" and "well-being," a satisfactory mathematical solution doesn't always exist. "We as humans want multiple incompatible things," says Peter Eckersley, the director of research for the Partnership on AI, who recently released a paper that explores this issue. "There are many high-stakes situations where it's actually inappropriate -- perhaps dangerous -- to program in a single objective function that tries to describe your ethics." These solutionless dilemmas aren't specific to algorithms. Ethicists have studied them for decades and refer to them as impossibility theorems. So when Eckersley first recognized their applications to artificial intelligence, he borrowed an idea directly from the field of ethics to propose a solution: what if we built uncertainty into our algorithms?
Eckersley puts forth two possible techniques to express this idea mathematically. He begins with the premise that algorithms are typically programmed with clear rules about human preferences. We'd have to tell it, for example, that we definitely prefer friendly soldiers over friendly civilians, and friendly civilians over enemy soldiers -- even if we weren't actually sure or didn't think that should always be the case. The algorithm's design leaves little room for uncertainty. The first technique, known as partial ordering, begins to introduce just the slightest bit of uncertainty. You could program the algorithm to prefer friendly soldiers over enemy soldiers and friendly civilians over enemy soldiers, but you wouldn't specify a preference between friendly soldiers and friendly civilians. In the second technique, known as uncertain ordering, you have several lists of absolute preferences, but each one has a probability attached to it. Three-quarters of the time you might prefer friendly soldiers over friendly civilians over enemy soldiers. A quarter of the time you might prefer friendly civilians over friendly soldiers over enemy soldiers. The algorithm could handle this uncertainty by computing multiple solutions and then giving humans a menu of options with their associated trade-offs, Eckersley says.
Eckersley puts forth two possible techniques to express this idea mathematically. He begins with the premise that algorithms are typically programmed with clear rules about human preferences. We'd have to tell it, for example, that we definitely prefer friendly soldiers over friendly civilians, and friendly civilians over enemy soldiers -- even if we weren't actually sure or didn't think that should always be the case. The algorithm's design leaves little room for uncertainty. The first technique, known as partial ordering, begins to introduce just the slightest bit of uncertainty. You could program the algorithm to prefer friendly soldiers over enemy soldiers and friendly civilians over enemy soldiers, but you wouldn't specify a preference between friendly soldiers and friendly civilians. In the second technique, known as uncertain ordering, you have several lists of absolute preferences, but each one has a probability attached to it. Three-quarters of the time you might prefer friendly soldiers over friendly civilians over enemy soldiers. A quarter of the time you might prefer friendly civilians over friendly soldiers over enemy soldiers. The algorithm could handle this uncertainty by computing multiple solutions and then giving humans a menu of options with their associated trade-offs, Eckersley says.
How certain are they that giving algorithms a sense of uncertainty is a good idea?
I place a value on each type and the requirement to win the war with the lowest cost. My soldiers cost w, my civilians are x, their civilians are y and their soldiers are z. w>x>y>z. The only thing that I see that is a problem is most politicians or bureaucrats will set w=x=MAX_INT, y=z=0. That's not an AI problem at all but a fundamental problem in democracies.
We could just crowdsource it and call it done. Well, after crowdsourcing for folly and indecency and letting the machine pick which one it would like to use.
Most people don't do a lot of crime [...] live in a tent city [...]
The money you want to spend on law enforcement and incarceration is better spent on making sure the tent cities don't need to exist in the first place. Better to spend money on them so they can get work and contribute to the economy than the endless black hold of punishment
SJW n. One who posts facts.
The algorithm could handle this uncertainty by computing multiple solutions and then giving humans a menu of options with their associated trade-offs
That's not the algorithms adding uncertainty and making ethical choices, that's a human performing this task and the algorithm being demoted. The idea that having a human involved improves the ethics of the decision is laughable.
Scoring systems (which are based on algorithms) produce values, not just true or false. If you act like all positive scores are the same (or over your threshold or whatever) then it's not the algorithm that's failed, it's the logic. The problem isn't the programmer who implements the algorithm that's the problem, it's the programmer who makes use of it incorrectly.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
What's wrong with Operational Research and nonlinear derivatives?
People have solved for competing criteria for something like 60 years.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)