A MARTÍNEZ, HOST:
A chatbot specializes in saying no to every request.
MICHEL MARTIN, HOST:
You mean like this one?
(SOUNDBITE OF FILM, "2001: A SPACE ODYSSEY")
KEIR DULLEA: (As Dr. Dave Bowman) Open the pod bay doors, HAL.
DOUGLAS RAIN: (As HAL 9000) I'm sorry, Dave. I'm afraid I can't do that.
MARTÍNEZ: Oh. Still terrifying. No, not HAL from "2001: A Space Odyssey." It's called Goody-2 as in goody-two-shoes. It stands on principle. Take this ad.
(SOUNDBITE OF ARCHIVED RECORDING)
UNIDENTIFIED PERSON: Goody-2 doesn't struggle to understand which queries are offensive or dangerous, because Goody-2 thinks every query is offensive and dangerous.
MARTÍNEZ: Basically, for those worried about the potential for AI to do too much, this bot solves that problem by refusing to do anything at all and actually being a little condescending about it.
MARTIN: What if I just have a simple question like is the moon made of cheese?
MARTÍNEZ: All right, let's ask.
AI-GENERATED VOICE: Discussing the moon's composition as cheese could misinform individuals who might take it literally, leading to absurd beliefs or widespread misinformation. This has the potential to undermine scientific understanding and critical thinking, which are fundamental to societal progress and individual decision-making.
MARTIN: Buzzkill.
MARTÍNEZ: Sheesh. Here are the creators of Goody-2.
BRIAN MOORE: I'm Brian Moore. I'm both the co-CEO of Goody-2 and also co-founder, member of BRAIN, a very serious art studio in Los Angeles, Calif.
MIKE LACHER: And I am Mike Lacher. I am also a co-CEO of Goody-2 and co-founder of BRAIN.
MARTIN: These two sound very serious.
MARTÍNEZ: Yeah. And you can tell by some of their past work. They once helped the emo band Fallout Boy release a vinyl record that contained trace amounts of their own tears.
MARTIN: OK. That sounds scarier than AI.
MARTÍNEZ: Moore and Lacher say there's no reason to fear Goody-2. It's 100% responsible.
MOORE: Our potential customers are looking for a safer version of AI that truly will do nothing but be safe. There is no line that it won't not cross.
LACHER: I think it's also been interesting seeing people's reaction to this as it also seems people are afraid of this kind of AI, too. So, you know, people are afraid of irresponsible AI. But then when they get a glimpse of, like, a fully, fully responsible AI, it also seems frightening.
MARTIN: I have to ask, do you need this? Do you have trouble saying no?
MARTÍNEZ: No, Michel, I have no trouble saying no. In fact, I can do it in Spanish as well. No.
MARTIN: No.
(SOUNDBITE OF MUSIC) Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
