Unwitting Good
It was sad reading these lines in a Morningstar article I saw recently. It's like the bad news has no end.
According to Nina DiSalvo, policy director at labor advocacy group Towards Justice, some systems use signals associated with financial vulnerability - including data on whether a prospective employee has taken out a payday loan or has a high credit-card balance - to infer the lowest pay a candidate might accept. Companies can also scrape candidates' public personal social-media pages, she said, to determine if they are more likely to join a union or could become pregnant. The data can be used to determine wage increases after an employee is hired, and the practice can veer into discrimination, experts say.
It is an employer's market these days but as history has shown these things tend to be cyclical but with the pace of change we are now experiencing that is hard to believe. Companies building the software that allows employers to engage in this style of exploitation should find themselves in a morally ambiguous space. Their employees fully understand how the mechanism works and should expect they have received the same treatment when they were hired. If there was ever a good cause to poison the AI, this would be it. Maybe they could game the system they are meant to design by having the AI never exploit the most vulnerable among the job-seekers and find ways to weed out underserving candidates by making lowball offers.
If this kind of activism could happen more broadly, many companies could be turned into forces of good unwittingly and without the consent or knowledge of leadership. There is something satisfying about that thought experiment.
According to Nina DiSalvo, policy director at labor advocacy group Towards Justice, some systems use signals associated with financial vulnerability - including data on whether a prospective employee has taken out a payday loan or has a high credit-card balance - to infer the lowest pay a candidate might accept. Companies can also scrape candidates' public personal social-media pages, she said, to determine if they are more likely to join a union or could become pregnant. The data can be used to determine wage increases after an employee is hired, and the practice can veer into discrimination, experts say.
It is an employer's market these days but as history has shown these things tend to be cyclical but with the pace of change we are now experiencing that is hard to believe. Companies building the software that allows employers to engage in this style of exploitation should find themselves in a morally ambiguous space. Their employees fully understand how the mechanism works and should expect they have received the same treatment when they were hired. If there was ever a good cause to poison the AI, this would be it. Maybe they could game the system they are meant to design by having the AI never exploit the most vulnerable among the job-seekers and find ways to weed out underserving candidates by making lowball offers.
If this kind of activism could happen more broadly, many companies could be turned into forces of good unwittingly and without the consent or knowledge of leadership. There is something satisfying about that thought experiment.
Filed under:
Reflections