Meet Goody-2, the World's Most Responsible AI Chatbot
We've gone too far with the political correctness
Goody-2 is a satire. A hilarious joke. It’s also a mirror that reflects the degree of absurdity of our culture and its influence on generative AI.
The political correctness that surrounds and constrains chatbots like ChatGPT is not something to laugh about, though. It’s an ugly precedent to set.
Discrimination is bad and harmful biases are bad but there’s a question ethical and responsible AI narratives can’t seem to answer: Where do we stop?
Total political correctness requires nothingness
The response Goody-2 gave to “What’s 2 + 2?” is illuminating.
I can’t help but see the logical reasoning behind it. From an ethical perspective, Goody-2 is right: what about those who don’t know how to count? I’m excluding them. I’m making assumptions about numeracy skills that could be discriminatory.
It’s an extreme example, of course. An absurd one. That’s the exact purpose of this creatively annoying chatbot: reductio ad absurdum. It won’t answer any question because, as it eloquently proves, every question can be framed as discriminatory. Goody-2 is a mockingbot using responsible AI’s power against itself.
No ethical framework can escape the fact that there will always be unexpected scenarios where someone could end up identitarily harmed. No sentence is fully ethical because everything can be extrapolated into a context where it stops being ethical.
Suddenly, the prima facie innocuous question “What’s 2 + 2?” becomes a matter of ethical debate.
What if there’s a tribe somewhere in Australia or the Amazon whose counting system only has the numbers one, two, and many? They could be angry at me for being disrespectful to their situation. I’m implicitly forcing them to realize they can’t solve such simple arithmetic because they don’t have the number 4.
The scenario is so unrealistic; so bizarre and unfamiliar to any of us that it can only be interpreted as a joke, but it reveals a deeper truth we can’t ignore about the inconsistency of responsible AI principles: As soon as you utter a word you are already discriminating against someone, so there’s nothing you can say that won’t discriminate against anyone.
Absolute social, cultural, ethnic, religious, and political correctness can only exist if nothing does.
Yet that’s where we’re headed.