4 ms·

Should AI software that’s smart enough to trick humans be forced to disclose itself. Google executives don’t have a clear answer yet. I would think that Google

Should AI software that’s smart enough to trick humans be forced to disclose itself. Google executives don’t have a clear answer yet. I would think that Google's executives shouldn't have a say in that question at all. As the chosen representatives of the people, that is a job for the Government.

rbanffy 8y ago
Wouldn't it be racist to force an intelligent entity to disclose its nature? Why does it matter if the person chatting with you on the phone is made of meat or a super-being who's having thousands of similar conversations at the same time?
- mdekkers 8y ago
  Wouldn't it be racist to force an intelligent entity to disclose its nature? This argument is hilarious, but I'll play along. First of all "racism" is about superiority, not about differentiation per-se. There is nothing inherently wrong about acknowledging differences, it is wrong when it becomes the sole reason for assuming superiority. That is the textbook definition of racism, so to answer your question: No, in my view it would not be wrong to force an intelligent entity that is not human to disclose this fact. It matters so that my approach to dealing with this "intelligent entity" can be adjusted accordingly, as per my needs. Why does it matter if the person chatting with you on the phone is made of meat or a super-being who's having thousands of similar conversations at the same time? It matters because today we do not yet understand anything about the drivers and motivations of this "super-being" and before we are clear on that, I would like to know who or what I am conversing with. I am pretty sure you are referring to "Minds" from "The Culture" series, which a) is fiction, and more importantly for the sake of this discussion, b) Minds have a relatively well understood context of motivations, morality and history of mostly benign interaction with Culture citizens. The "Robocall version 2.0" touted by Google doesn't have that, and given Googles past track record, they don't really give a flying fuck about the general population. When it comes to the proprietor and owner of this technology today, we have been comprehensively identified as "the product" instead of "the customer" -- it is my right as a private citizen to know who or what I am interacting with, especially if it is a Google product. Finally, whereas dealing with other humans can be understood on a base level of expected intelligence and wits, as well as shared identity, history and culture, interaction with "super-being" would by its' very definition be an unequal exchange. I have every right to know if my interlocutor had any kind of extreme advantage over me (you know, because it is a "super-being"). Or am I being speceist now?
  - rbanffy 8y ago
    > First of all "racism" is about superiority, not about differentiation per-se. In this case, the IA would disclose its nature and its lack of independent agency. It'd disclose the fact it's not considered a living being and not afforded any rights. The human, then, could engage in all sorts of abuse (and Microsoft had a lovely chatbot that had its "brain" damaged by trolls) that could leave lasting effects on its persisted state. > I am pretty sure you are referring to "Minds" from "The Culture" series You are completely wrong. Google's "Robocall 2.0" is nowhere near close a sci-fi worthy AI.

rbanffy 8y ago

Wouldn't it be racist to force an intelligent entity to disclose its nature? Why does it matter if the person chatting with you on the phone is made of meat or a super-being who's having thousands of similar conversations at the same time?

mdekkers 8y ago
Wouldn't it be racist to force an intelligent entity to disclose its nature? This argument is hilarious, but I'll play along. First of all "racism" is about superiority, not about differentiation per-se. There is nothing inherently wrong about acknowledging differences, it is wrong when it becomes the sole reason for assuming superiority. That is the textbook definition of racism, so to answer your question: No, in my view it would not be wrong to force an intelligent entity that is not human to disclose this fact. It matters so that my approach to dealing with this "intelligent entity" can be adjusted accordingly, as per my needs. Why does it matter if the person chatting with you on the phone is made of meat or a super-being who's having thousands of similar conversations at the same time? It matters because today we do not yet understand anything about the drivers and motivations of this "super-being" and before we are clear on that, I would like to know who or what I am conversing with. I am pretty sure you are referring to "Minds" from "The Culture" series, which a) is fiction, and more importantly for the sake of this discussion, b) Minds have a relatively well understood context of motivations, morality and history of mostly benign interaction with Culture citizens. The "Robocall version 2.0" touted by Google doesn't have that, and given Googles past track record, they don't really give a flying fuck about the general population. When it comes to the proprietor and owner of this technology today, we have been comprehensively identified as "the product" instead of "the customer" -- it is my right as a private citizen to know who or what I am interacting with, especially if it is a Google product. Finally, whereas dealing with other humans can be understood on a base level of expected intelligence and wits, as well as shared identity, history and culture, interaction with "super-being" would by its' very definition be an unequal exchange. I have every right to know if my interlocutor had any kind of extreme advantage over me (you know, because it is a "super-being"). Or am I being speceist now?
- rbanffy 8y ago
  > First of all "racism" is about superiority, not about differentiation per-se. In this case, the IA would disclose its nature and its lack of independent agency. It'd disclose the fact it's not considered a living being and not afforded any rights. The human, then, could engage in all sorts of abuse (and Microsoft had a lovely chatbot that had its "brain" damaged by trolls) that could leave lasting effects on its persisted state. > I am pretty sure you are referring to "Minds" from "The Culture" series You are completely wrong. Google's "Robocall 2.0" is nowhere near close a sci-fi worthy AI.