[ivan.moony]

A day after Microsoft introduced an innocent Artificial Intelligence chat robot to Twitter it has had to delete it after it transformed into an evil Hitler-loving, incestual sex-promoting, 'Bush did 9/11'-proclaiming robot.

More here...

Should we be worried?

[Freddy]

More amused than worried I think 

[Don Patrick]

You should only be worried if you were planning to let your A.I. learn from the internet unsupervised too.

[ivan.moony]

Don't you know it? I plan just that: unsupervised crowdsourced learning.

And I have a solution for that: ethical filter for learnt stuff. It should drill those funny twitter 'mericans, if I calculated correctly.

[DemonRaven]

Any chatbot creator could have told them that no matter what they have to be monitored. Like it or not you can not be lazy with them if you are going to let other people "train" them.

[ivan.moony]

If I calculated correctly, with true AI there is no jokes, no matter how funny those twitters might look like. If it is supposed to be 100 times smarter than a human, imagine what it can do in crime, especially if AI is replicated numerous times.

I've lived through one war in my lifetime in the place I live and I can tell you: a war is not a joke. I rejected my AI ideas numerous times in my whole work and planning, but every time I have found a solution to possible mistakes. Right now I'm still in, considering all heard and learnt by now, gathered over some 15 years.

I'm particularly interested in electromagnetic impulse bombs that would instantly disable all chips in some radius if anything goes wrong. Otherwise, I'm interested in AI rejecting bad ideas and in AI proposing solutions over AI abuse. It is not all honey and daisies, but I think that humanity can work it out if we take it serious enough. We have much to gain.

Microsoft offered half-solution parrot machine and started yelling: "People, this is AI, buy while you can". They promised too much and encountered disgrace because they weren't serious enough. And possible war is more than serious, especially if it can come from machines. Considering anything less should be illegal, because the danger could be real.

Edit: Some of us could get near constructing true AI soon. Do you think "don't worry, be happy" is the right message we want to send to each other?

[Art]

I'm particularly interested in electromagnetic impulse bombs that would instantly disable all chips in some radius if anything goes wrong. 

Faraday Cage.

[Data]

Mr x
I've got a great idea - lets make a chatbot that learns from conversations, then get some teenagers to teach it for us.

Mr Y
Wouldn't it be better to get scientists, lecturers or people with more life experience to teach the bot ?

Mr x
No, teenagers will be fine, what could possibly go wrong ?

Mr Y
Sounds good 

[ivan.moony]

I'm particularly interested in electromagnetic impulse bombs that would instantly disable all chips in some radius if anything goes wrong. 

Faraday Cage.

... Protects contents in some extent... Which means you have to be very close to have an impact of enough power. Faraday cage certainly complicates things. But there is other thing too... AI could be a scientist, inventing something new, like optical processors. Anyway, if something goes wrong, it will be no good.

[Freddy]

Data, I bet that's pretty much how it went