MSFT had to turn off its twitter bot .....

[Joe Fan]

.... after only 15 hours
Kind of a funny story
[Trigger warning for some even more offensive tweets in the article, worst than what is visible below]
http://www.businessinsider.com/micr...l-tweets-from-ai-chatbot-tay-2016-3?r=UK&IR=T

 

[UTChE96]

Bizarre but hilarious.

 

[Seattle Husker]

MSFT was just reminded that the internet is a dark place.

 

[chango]

Joe Fan - you are the Horfans Twitter Guru, hands down!

 

[MudHorn]

From the linked article in @Joe Fan's post:

It's clear that Microsoft's developers didn't include any filters on what words Tay could or could not use.

Good job, Microsoft.

 

[Seattle Husker]

The article points to this project as a "Machine Learning exercise". Did the internet just teach "Tay" to be a bigoted racist?

 

[Htown77]

Based on the average youtube or news article comment sections, we may need to worry about accidentally creating nazi artificial inteligence.

Microsoft should use hornfans instead. Any hornfans based AI will not discriminate based on religion, color or creed.... but I cannot promise that it would look favorably upon sooner or aggy.

 

[Driver 8]

This is what happens when Microsoft tries to be edgy and hip. They still have Clippy in their DNA.

 

[Phil Elliott]

They still have Clippy in their DNA.

Good one!