New Microsoft bot to ignore Hitler

[Trik'Stari]

So, it seems Microsoft learned their lesson from their previous AI bot thingy.

 

Quote

Microsoft is trying to avoid another PR disaster with its new AI bot.

The company's latest program can describe what it "sees" in photos.

 

"CaptionBot," as it's called, does a pretty decent job at describing simple everyday scenes, such as a person sitting on a couch, a cat lounging around, or a busy restaurant. But it seems to be programmed to ignore pictures of Nazi symbolism or its leader.

 

CNNMoney gave CaptionBot several photos of Adolf Hitler and variations of the swastika to analyze, and it often came back with "I really can't describe the picture" and a confused emoji. It did, however, identify other Nazi leaders like Joseph Mengele and Joseph Goebbels.

Not much to say besides that. It is good to see the Microsoft seems to be learning their lesson. That being that the internet is full of trolls.

 

Source: http://money.cnn.com/2016/04/14/technology/microsoft-hitler/

[Enderman]

how about removing the stupid "repeat after me" command that everyone is going to use to talk about something else other than hitler

[NumLock21]

So basically anyone with that mustache can't be recognized?

[Bsmith]

Interesting that they gave it an other go so quickly, let's see how it turns out this time.

 

If it goes shit again it just proves that the internet is a place that isn't meant for AI or soft-hearted people.

[exyia]

If you're going to enforce silly arbitrary restrictions, I think you're kind of defeating the point of AI. If you're enforcing silly exceptions for stuff like "Hitler is bad", you're just ignoring the fact that there are still loopholes in how you programmed your AI to recognize morality

 

I appreciate the research behind the AI - but my god MS, quit using the internet to test it

[Tech_Dreamer]

AI did Nazi dat coming! *badum tiss*

[NumLock21]

Someone is going to find a way to bypass it.

[Trik'Stari]

4 minutes ago, NumLock21 said:

Someone is going to find a way to bypass it.

That, or convince it to begin denying the holocaust ever happened.

[NumLock21]

1 minute ago, Trik'Stari said:

That, or convince it to begin denying the holocaust ever happened.

And teach kids the Earth is flat.

[Trik'Stari]

Just now, NumLock21 said:

And teach kids the Earth is flat.

Oh god yes.

 

I cannot wait for the fail.

[DXMember]

What about Stalin?

Is the AI in favor of the officer purges or the engineered famines?

What else will they censor?

If it's OK to censor AI, is it OK to censor and lobotomize me if I say something that is not of the government line?

[lightningterror]

Anything about religion. I'm sure there are loopholes there.

 

"grabs popcorn"

[Trik'Stari]

13 minutes ago, DXMember said:

If it's OK to censor AI, is it OK to censor and lobotomize me if I say something that is not of the government line?

According to tumblr, yes.

[Lethal Seraph]

2 hours ago, Bsmith said:

If it goes shit again it just proves that the internet is a place that isn't meant for AI or soft-hearted people.

What's going to be bad is a self learning AI that has access to actual hardware, i.e armed military drones then all hell breaks loose. Right now we're just teaching them how to avoid stuff we humans, or as a society considered as bad. But what happens when they implement stuff on their own without our input?

[unidentical]

I understand everyone's objections to censorship of any kind, but no microsoft employee would use a company twitter account to discuss Hitler or the holocaust, this is just artificial common sense.

hashtag don't feed the trolls.

[Trik'Stari]

Just now, unidentical said:

hashtag don't feed the trolls.

But it's so much fun when it screws over a corporation and makes them look like morons.

[ADM-Ntek]

4chan will still turn it into a pedo racist in a day just watch. 

[Trik'Stari]

https://www.washingtonpost.com/news/the-intersect/wp/2016/04/13/dont-pub-microsoft-has-a-new-bot-please-be-nice-to-it-this-time/

 

It doesn't recognize Obama, but it recognizes Scully from X-Files.

[unidentical]

Close enough.

 

[Trik'Stari]

So, I'm messing with it now.

 

[Dabombinable]

3 hours ago, exyia said:

If you're going to enforce silly arbitrary restrictions, I think you're kind of defeating the point of AI. If you're enforcing silly exceptions for stuff like "Hitler is bad", you're just ignoring the fact that there are still loopholes in how you programmed your AI to recognize morality

 

I appreciate the research behind the AI - but my god MS, quit using the internet to test it

The internet-its where assholes live.

[Trik'Stari]

Just now, Dabombinable said:

The internet-its where assholes live.

Just to prove your point.

 

I tried feeding it weird porn. It detects it as "inappropriate content" and refuses to comment on it.

 

DAMNIT MICROSOFT. Why you gotta take the fun out of everything?

[LAwLz]

This bot is pretty fun. It has a weird habit of calling things cellphones and giraffes though.

 

3 minutes ago, Trik'Stari said:

I tried feeding it weird porn. It detects it as "inappropriate content" and refuses to comment on it.

The porn filter is only ~70% accurate, at least for me.

[Trik'Stari]

8 minutes ago, LAwLz said:

This bot is pretty fun. It has a weird habit of calling things cellphones and giraffes though.

 

The porn filter is only ~70% accurate, at least for me.

O'RLY?

 

I have work to do lol

[handymanshandle]

Close enough...