4 ms·

It's an -ism a little like sexism, racism, ageism and disability discrimination. There will be countless instances that will be hard to prove the loss - as you

It's an -ism a little like sexism, racism, ageism and disability discrimination. There will be countless instances that will be hard to prove the loss - as you don't have transparency of what the other side is thinking/saying. Whether it's some dubious data from suspect source, or it's your gender, you might find some opportunities more difficult, premiums more expensive, interviews not bringing job offer. When it's just one isolated, it proves little, but in aggregate... We already know many of these data models are very wrong. When Amazon recommend products for me, they do so terribly. When Google and Facebook famously opened up their respective "what we know about you" pages a few years back, they were - for most in our office - dramatically and hilariously wrong. Often getting every key point wrong like gender, age group, and 9 of 10 interests. AI models are often reported to have gained or inbuilt bias and racism. It doesn't feel obviously bad, but in the cargo cult of targeted advertising, they've demonstrated they're unfit for purpose. Yet the vast data stores influence all our lives daily -- we really need an audit to identify all those micro impacts and data derived discrimination, as it's undermining any belief in a rational basis and a fair chance in society.

criddell 7y ago
It seems when I read arguments like this it's always about what I lose. For every job that I lose somebody else wins perhaps also because of data gleaned from surveillance. It isn't necessarily negative. The insurance argument is pretty persuasive though. Effective insurance requires a big pool. For the same reason that ultra-precise categorization for insurance ruins insurance, ultra-precise targeting of individuals ruins advertising if we want ad-supported platforms. Regulation is more important than ever.
- NeedMoreTea 7y ago
  Thing is, I want a level playing field whether applying for jobs, insurance, health, you name it. A hidden bias is still a bias, and particularly hidden as so much machine processing and AI is a black box unknown even to the creators -- it just spits out (hopefully) helpful results. I don't want a chilling effect on the next generation such that every party or moment of drunkenness can be held against them in job search. Every teen has moments of idiocy -- mine thankfully were not in the age of constant surveillance, and there wasn't a smart phone or CCTV in sight. Not much freedom to grow in a world where every worst moment goes viral, or gets you on one of the "next door" neighbourhood snooping apps. Where hanging around with friends on a street corner has old folks flagging suspicious behaviour. Give us all room to breathe. Interfere via authorities when you have genuine reason to suspect. No, I don't much like where we're headed and preferred the proposition before. Totally agree on regulation, but it seems to be coming far too slowly -- globally.

criddell 7y ago

It seems when I read arguments like this it's always about what I lose. For every job that I lose somebody else wins perhaps also because of data gleaned from surveillance. It isn't necessarily negative. The insurance argument is pretty persuasive though. Effective insurance requires a big pool. For the same reason that ultra-precise categorization for insurance ruins insurance, ultra-precise targeting of individuals ruins advertising if we want ad-supported platforms. Regulation is more important than ever.

NeedMoreTea 7y ago
Thing is, I want a level playing field whether applying for jobs, insurance, health, you name it. A hidden bias is still a bias, and particularly hidden as so much machine processing and AI is a black box unknown even to the creators -- it just spits out (hopefully) helpful results. I don't want a chilling effect on the next generation such that every party or moment of drunkenness can be held against them in job search. Every teen has moments of idiocy -- mine thankfully were not in the age of constant surveillance, and there wasn't a smart phone or CCTV in sight. Not much freedom to grow in a world where every worst moment goes viral, or gets you on one of the "next door" neighbourhood snooping apps. Where hanging around with friends on a street corner has old folks flagging suspicious behaviour. Give us all room to breathe. Interfere via authorities when you have genuine reason to suspect. No, I don't much like where we're headed and preferred the proposition before. Totally agree on regulation, but it seems to be coming far too slowly -- globally.