Filtered for monkeys and A.I.
14.56, Thursday 8 Jan 2015 Link to this post
There was an interesting dispute last year around that photo the monkey took of itself… or rather, a photographer was out to take pictures, but a monkey nicked the camera and took loads of pictures, then the photographer picked out a particularly good one and shared it, and everyone called it the “monkey selfie.”
Then Wikipedia published the photo without asking the photographer, and justified it by saying that actually the monkey owned the photo.
Anne takes this to an interesting place:
Does the monkey have agency? Clearly.
Is the monkey the author-photographer? Sure.
Is the monkey the owner? Possibly.
And if this nonhuman has agency, and the power of authorship and ownership, what about other nonhumans?
What about the camera? What kind of agency does it have? Can a camera author an image? Own a photograph?
Is the camera the author of the photo?
Amazon’s robotic fulfilment army [video].
In other robot news, If Your Robot Buys Illegal Drugs, Have You Committed a Crime?
If I give you a cake, did I deliberately make you happy? Of course.
If I order it to be sent online? Of course.
If I flip a coin and if it comes up heads then you get cake, then I flip the coin and it’s heads? Yes.
If I write a program that can buy all kinds of stuff and sometimes buys cake? Probably, yes.
If I write an artificial intelligence that wants to delight you, and it happens to get you cake, is that still my agency? I think so, yes.
At a certain point of complexity (is it complexity that matters?) it’s no longer my agency. But I don’t know what that is.
I want there to be like an age of majority for systems of cause and effect, with a threshold of complexity rather than years since birth.
x.ai is an automated, email personal assistant to help you arrange meetings. It uses artificial intelligence.
Matt Turck’s analysis of A.I. as a space is great. How artificial intelligence will come to market:
we’re about to witness the emergence of a number of deeply focused AI-powered applications that will achieve commercial success by solving in a definitive manner very specific issues.
…which is a great way to think about it. I don’t care how this personal assistant works, but it’s great that the company behind it were able to create it. And they were able to create it because of A.I.
If it does something weird or wrong, whose fault is that? The programmers? In some cases. But in most cases, if the person I’m arranging a meeting with is let down or offended or pissed off… it’ll be my fault, the user, the person who introduced the A.I. into the situation.
So how do I “interview” an A.I. for this personal assistant job role? How do I get to know them, assure myself they’re not going to do anything bonkers?
Intel have released a computer that plugs into the back of a TV.