We have rather learnt that intelligence, the capacity for it at least, is at least somewhat innate. But how that is expressed is a matter of the environment that intelligence is trained in. People brought up in a Catholic society tend to be Catholics, people brought up in a fascist one fascists, communist communist and so on. Sure, there are always exceptions and rejections of then received wisdom and so on but it is pretty obvious that the environment matters for how the intelligence is trained.
Which brings us to the, err, idiots, at Microsoft:
Tay, Microsoft Corp’s so-called chatbot that uses artificial intelligence to engage with millennials on Twitter, lasted less than a day before it was hobbled by a barrage of racist and sexist comments by Twitter users that it parroted back to them.
TayTweets (@TayandYou), which began tweeting on Wednesday, was designed to become “smarter” as more users interacted with it, according to its Twitter biography. But it was shut down by Microsoft early on Thursday after it made a series of inappropriate tweets.
A Microsoft representative said on Thursday that the company was “making adjustments” to the chatbot while the account is quiet.
“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,” the representative said in a written statement supplied to Reuters, without elaborating.
The Trollerati looked at this, saw that the AI could be trained, saw that this was good and thus trained it to be a racist buffoon.
Who thought it would be any different?
There’s always the drunk Uncle* who teaches the two year old nephew to say “Fuck!” in order to shock the mothers’ friends. The training needs to be done by those who actually care about the outcome, not those looking for a larff.
*Perhaps not Uncle but someone, somewhere,