3 ms·

Sorry, but I don't see a single thing of value here but more sensationalism from the purveyors who've been at this for years. People like the authors Gebru and

Sorry, but I don't see a single thing of value here but more sensationalism from the purveyors who've been at this for years. People like the authors Gebru and Mitchell, and now Lemoine have contributed negatively to the field while continuing to tarnish the domain with their obviously biased views that serve their own agendas. Lemoine now claims to be a "mystic Christian priest" and that position is somehow supposed to validate his wild rants. Yes, the current state of AI is biased because the training and sample data themselves are not unbiased. This is the reality, and there's work to be done. However, with their increasingly outlandish rants ("This AI is a person") [1] these people are just looking at riding the AI bandwagon to serve their own twisted agendas, i.e., make a lot of noise and hope to get paid to keep doing that. [1] https://www.wired.com/story/blake-lemoine-google-lamda-ai-bigotry/ https://www.wired.com/story/blake-lemoine-google-lamda-ai-bi...

hatenberg 4y ago
Nice ad hominem you got there.
abrichr 4y ago
This seems valuable to me: > Scientists and engineers should focus on building models that meet people’s needs for different tasks, and that can be evaluated on that basis, rather than claiming they’re creating über intelligence. Similarly, we urge the media to focus on holding power to account, rather than falling for the bedazzlement of seemingly magical AI systems, hyped by corporations that benefit from misleading the public as to what these products actually are.
zzzeek 4y ago
bet: in five years there will be legal cases attempting to argue that AIs can count as legal persons, which will be for the purposes of absolving a corporation being sued for harm from responsibility for those acts as "the AI was acting independently". That is by far the most important and very plausible case Gebru is making and we will definitely see it happen soon. "Blame the AI" would be an incredibly easy way to absolve bad actors of responsibility if they could clear such a legal hurdle and all of the "AI is sentient" hype most certainly leads towards that outcome .

hatenberg 4y ago

Nice ad hominem you got there.

abrichr 4y ago

This seems valuable to me: > Scientists and engineers should focus on building models that meet people’s needs for different tasks, and that can be evaluated on that basis, rather than claiming they’re creating über intelligence. Similarly, we urge the media to focus on holding power to account, rather than falling for the bedazzlement of seemingly magical AI systems, hyped by corporations that benefit from misleading the public as to what these products actually are.

zzzeek 4y ago

bet: in five years there will be legal cases attempting to argue that AIs can count as legal persons, which will be for the purposes of absolving a corporation being sued for harm from responsibility for those acts as "the AI was acting independently". That is by far the most important and very plausible case Gebru is making and we will definitely see it happen soon. "Blame the AI" would be an incredibly easy way to absolve bad actors of responsibility if they could clear such a legal hurdle and all of the "AI is sentient" hype most certainly leads towards that outcome .