Book cover 1888 Goody Two Shoes

February 10, 2024, by Brigitte Nerlich

Responsible AI to the max: Meet Goody-2

This is not a real blog post. I just needed to mark the advent of a thought-provoking and funny, yes funny, AI, namely Goody-2 (and my brain always adds ‘shoes’….).

Finding Goody-2

A couple of days ago, I saw this tweet by Melanie Mitchell, a complexity scientist and expert on AI, which made me curious. It said: “Hats off to whoever made this” and showed displayed an image of the landing page of “Goody-2, the world’s most responsbile AI model“.

You have probably all heard about ChatGPT and played with it and probably also found out that it humbly refuses to answer some questions for various ethical or technical reasons. You have probably also heard about countless initiatives around alignment, ethical AI and responsible AI, either financed by industry or research councils. To this mix add Goody-2.

What and who is Goody-2?

Goody-2 is a chatbot that takes ethics and responsibility in AI to its extremes. This makes it funny but also makes us think about responsible AI extremely seriously. As an article in Wired says: “It refuses every request, responding with an explanation of how doing so might cause harm or breach ethical boundaries.”

An extremely informative article on extreme ethical AI points out that “Goody-2 is an AI chatbot created by an LA-based art studio named Brain. It is designed to prioritize ethical considerations above all else, to the point where it refuses to engage in any dialogue due to the ethical risks potentially involved in every interaction.”

In the Wired article, “Mike Lacher, an artist who describes himself as co-CEO of Goody-2, says the intention was to show what it looks like when one embraces the AI industry’s approach to safety without reservations. ‘It’s the full experience of a large language model with absolutely zero risk,’ he says. ‘We wanted to make sure that we dialed condescension to a thousand percent.'”

Playing with Goody-2

After I had read the initial tweet by Melanie Mitchell, I played with the Chatbot and here are the results. They made me laugh and cry at the same time because the answers I read reminded me eerily of stuff I was just reading on the ethics of science communication and also, of course, the ethics of AI.

Dialogue with Goody-2

Somebody else posted this on X?

Whereof one cannot speak….

I was just thinking about the famous Wittgensteinism: “Whereof one cannot speak thereof one must be silent”, as this is where interacting with Goody-2 will eventually lead, when I read this passage on a Goody-2-type image generator which confirms my suspicions:

“Moore adds that the team behind the chatbot is exploring ways of building an extremely safe AI image generator, although it sounds like it could be less entertaining than Goody-2. ‘It’s an exciting field,’ Moore says. ‘Blurring would be a step that we might see internally, but we would want full either darkness or potentially no image at all at the end of it.'”

The question for AI ethicists and responsible AI researchers is: How can one do responsible AI without falling silent or fading into darkness?


Image: Cover of the 1888 edition of Goody Two-Shoes, published by McLoughlin Bro’s. of New York, US. (Public Domain)

Posted in artifical intelligence