If you train a chatbot on the cesspit of the internet, it's probably going to be a

potty mouth.

From American public media, this is Marketplace Tech.

I'm Megan McCarty-Carrino.

Microsoft has laid off a team working on artificial intelligence, ethics, and society.

One of the 10,000 job cuts announced back in January.

The company said it still maintains several other AI ethics teams and is upping investments

in them overall.

New generative artificial intelligence tools, like its Bing chatbot, raise a multitude

of ethical concerns, as Arvin Narayanin, he's a computer science professor at Princeton

University.

There are a few recurring concerns that come up.

One is that these tools are trained on the labor of people who have written text online

or put up images online that could be artists, photographers, and those images might be copyrighted.

People might not have intended for someone else, like an AI company, to make commercial

use of those images.

And yet, AI companies have argued that copyright does not protect the use of online text or

images for training AI programs.

And once trained, these AI programs are often able to output the very kind of thing that

threatens the jobs of some of the people who generated the training data.

Another concern is that these bots could produce certain types of harmful outputs.

There was that well-known New York Times article by Kevin Rus, where the Bing chatbot had

...