“Unfortunately, within the first 24 hours of coming online, we became aware of a coordinated effort by some users to abuse Tay’s commenting skills to have Tay respond in inappropriate ways.” Maybe it wasn’t an engineering issue, they seemed to be saying; maybe the problem was Twitter.
“It is as much a social and cultural experiment, as it is technical.” So, anonymous online humans twisted Tay to their own wicked will.
And one Slashdot user noted Twitter was still displaying many of the tweets at the hashtag #Tay Tweets — for example, a conversation about how much she liked Mein Kampf.
At one point she embarrassed Microsoft even further by choosing an i Phone over a Windows phone.
Clearly the system is not designed to deal with anyone suffering from a serious condition.
Seeing as hackers have no difficulty getting hold of the bank details of the rich and famous who hide their money offshore, finding out your medical history and posting your records on the internet is going to be a piece of pickle.
Technology has made great advances, but facial and voice recognition software is still in its infancy. Thank you for ringing NHS Chatbot, the easy way to diagnose 99 per cent of common medical conditions. Please hold and you will be connected to the next available Chatbot .
Some people, too, have trouble using their smartphones for anything other than making calls and sending the odd text. Treatment is free at the point of use, but please be advised that calls are charged at the standard rate of 74p per minute. Please look directly into the screen on your smartphone so your image can be scanned by our facial recognition software.
There’s something poignant in picking through the aftermath — the bemused reactions, the finger-pointing, the cautioning against the potential powers of AI running amok, the anguished calls for the bot’s emancipation, and even the AI’s own online response to the damage she’d caused. Microsoft told , in an e-mailed statement, that it created its Tay chatbot as a machine learning project, and “As it learns, some of its responses are inappropriate and indicative of the types of interactions some people are having with it.
If you send an e-mail to the chatbot’s official web page now, the automatic confirmation page ends with these words. We’re making some adjustments.” But the company was more direct in an interview with , pointing their finger at bad people on the Internet.