Microsoft’s chatbot Tay learns how to be a racist from Twitter users

[Yo_Mama_Been_Loggin]

The internet can be an ugly place, which is why every parent knows you must monitor your child’s online interactions.

That is a lesson that Microsoft did not heed until after its freshly launched artificial intelligence Twitter chatbot Tay started parroting racist and sexist remarks from other tweets.

Less than a day after launching Tay, Microsoft has deleted all the chatbot’s messages, including tweets praising Hitler and genocide and tweets spouting hatred for African Americans.

Yea, it got pretty ugly.

Microsoft (Nasdaq: MSFT) unveiled Tay Wednesday as an experiment in artificial intelligence. Tay’s task was to display how it could master conversational understanding.

http://www.bizjournals.com/seattle/blog/techflash/2016/03/microsoft-s-chat-box-learns-how-to-be-a-racist.html?ana=e_sea_rdup&s=newsletter&ed=2016-03-24&u=ColXVN5SPzQtLHFP87ho2w07857290&t=1458835269&j=71710192

[0rganism]

1. seems like a successful experiment to me

they set up "Tay" to learn behavior from its social media peers, right? well, mission accomplished, eh.
now maybe they need to upgrade "Tay" with some basis for ethical discernment so it can appropriately weigh the input it receives. that is going to be challenging.

[TexasBushwhacker]

2. If its goal was to reveal just how racist and sexist

the US is, MISSION ACCOMPLISHED!

[B2G]

6. Only the US uses the internet?

Lolol. You're cute.

[ohnoyoudidnt]

3. What happens when you let Internet users teach a blank slate AI program?

Some of the programmers had to expect this outcome.

[FSogol]

4. Damn, the Klan will be out of work once the new racist AIs take over all the hate work on the net.

[LongTomH]

5. I've read articles by 'Singulatarians" on "Creating Friendly AI"

I wonder what their reaction to this 'chatbot' will be!