3 ms·

There is a distinction to be made here between brute utility and morality. If a person commits a crime and serves the time for it and lives beyond the point whe

by dsfsdfd 12y ago

There is a distinction to be made here between brute utility and morality. If a person commits a crime and serves the time for it and lives beyond the point where that crime has to be reported when applying for a job - should that previous conviction alter their likelihood of getting hired? Society as a whole may decide that it should not, yet the processes which determine the chances of hiring will not reflect the decision made by society as a whole. There will be a very large number of conflicts of this type, on the one hand businesses want to maximise profits and on the other society as a whole wants to impose certain moral and economic values upon those businesses. This is part of a more general tension between different structures within a society. The difference here is that before the conflict was less obvious because the agents involved could hide their motivation, when we are using machine agents for decision making we can easily query them and observe their bias objectively. So rather than creating discord between the various motives of society, it is really only making them more visible. Which I would say is a good thing.

blfr 12y ago
If people are fine with college records from 20+ years ago influencing the likelihood of getting hired but want companies to ignore criminal convictions, or ban IQ test, I'd rather go with a machine making the decisions. Current process seems purely political with massive rents extraction ($100k for a degree, loan guarantees), almost naked votes buying (subsidies, "free" plan for whatever). And moral arguments are used mostly for cover (non-dischargeable debts?). Perhaps we can sidestep those problems, or at least escape from them for a while.
- pron 12y ago
  > Perhaps we can sidestep those problems, or at least escape from them for a while. But we can't. Any decision we make -- even one based purely on data -- is a political one, because we can choose to change the data by changing society. If you say we should make future decisions (which affect lives and future data) based on current correlations only, you're executing a conservative political policy of maintaining the status quo. If you program computers to just take into account existing data (and not, say, desired future outcomes for society), you're programming conservative computers.
  - blfr 12y ago
    I'm pretty sure we can sidestep degree requirements and the associated mess with a better candidate screening technique. Why not? Also, the computer programs take into account the desired outcomes and the information they can obtain experimentally (through A/B testing for example), not just existing data.
    - eli_gottlieb 12y ago
      >I'm pretty sure we can sidestep degree requirements and the associated mess with a better candidate screening technique. Why not? Because feature-learning problems in the setting of nearly-unlimited sets of possible features are very hard.
- theorique 12y ago
  Part of knowing what college a person went to is a proxy for knowing their likely SAT scores which is a proxy for an (illegal) IQ test. So it's a moderately effective workaround. Of course, college can be gamed in a lot of different ways (family money, connections, hustle) that do not relate specifically to IQ or to job performance (and others that, of course, do).

by dsfsdfd 12y ago

blfr 12y ago

If people are fine with college records from 20+ years ago influencing the likelihood of getting hired but want companies to ignore criminal convictions, or ban IQ test, I'd rather go with a machine making the decisions. Current process seems purely political with massive rents extraction ($100k for a degree, loan guarantees), almost naked votes buying (subsidies, "free" plan for whatever). And moral arguments are used mostly for cover (non-dischargeable debts?). Perhaps we can sidestep those problems, or at least escape from them for a while.

pron 12y ago
> Perhaps we can sidestep those problems, or at least escape from them for a while. But we can't. Any decision we make -- even one based purely on data -- is a political one, because we can choose to change the data by changing society. If you say we should make future decisions (which affect lives and future data) based on current correlations only, you're executing a conservative political policy of maintaining the status quo. If you program computers to just take into account existing data (and not, say, desired future outcomes for society), you're programming conservative computers.
- blfr 12y ago
  I'm pretty sure we can sidestep degree requirements and the associated mess with a better candidate screening technique. Why not? Also, the computer programs take into account the desired outcomes and the information they can obtain experimentally (through A/B testing for example), not just existing data.
  - eli_gottlieb 12y ago
    >I'm pretty sure we can sidestep degree requirements and the associated mess with a better candidate screening technique. Why not? Because feature-learning problems in the setting of nearly-unlimited sets of possible features are very hard.
theorique 12y ago
Part of knowing what college a person went to is a proxy for knowing their likely SAT scores which is a proxy for an (illegal) IQ test. So it's a moderately effective workaround. Of course, college can be gamed in a lot of different ways (family money, connections, hustle) that do not relate specifically to IQ or to job performance (and others that, of course, do).