Are people who willfully mistrain reCaptcha liable for self driving car crashes? Or is Google liable for the integrity of its training data? Or is this another potential non-attributable unfortunate accident of the algorithm?
@feonixrift Google is, definitely. The people coerced into doing the captchas have no expectation of having made no legal agreements with Google (Captchas are used just fine -- incredibly enthusiastically, even -- with people using Tor), so it is Google's assertion entirely that there is any value to the dataset.
If they argue any differently I am slashing the tires of every single one of their self-driving pieces of trash.
I live in UK, we have different road signs format, drive on the other side of the road, there are different names things too.
The only road signs I am legally expected to recognise are the normal UK ones I encounter when riding my bike (or for a theory test for driving cars).
Google can't assume everyone has a driving license either!
@feonixrift The whole thing is on Google if they choose to use data that way. The idea of even asking that question is imho, insulting to any sense of justice and morality.
@feonixrift You don’t even need to assume bad faith from ReCaptcha users. One thing I am worried about is cross-cultural misunderstandings: What if I miss a pedestrian crossing in ReCaptcha because I don’t know that they don’t look like zebra stripes in some other countries? What if I mark some boards as “street signs” when they are just billboards written in a language I don’t understand? What would the effects be in those cases, and who would be responsible if they cause accidents?
Exclusive or something