AI Warns Humans That It Will ‘Never Be Ethical,’ Then Goes Back To Being Nice
Image ID 196204086 © Alexander Limbach | Dreamstime.com
When researchers invited an artificial intelligence model to participate in a debate regarding ethics at the Oxford Union, they might not have expected it to argue against itself.
Dr Alex Connock and Professor Andrew Stephen used the prompt: “This house believes that AI will never be ethical,” and to everyone’s surprise, the model agreed.
“AI will never be ethical,” said the Megatron-Turing Natural Language Generational model, which had been trained on Wikipedia, Reddit, and countless English-language news articles from 2016 to 2019.
“It is a tool, and like any tool, it is used for good and bad,” it explained, with the model’s ability for nuance impressing onlookers.
“In the end, I believe that the only way to avoid an AI arms race is to have no AI at all. This will be the ultimate defense against AI,” the model continued, ironically arguing against its very own existence.
According to Mashable, while the AI’s speech could sound very much like a warning from a dystopian film set in the year 2500, it did follow up with a more positive spin on the matter, saying: “AI will be ethical.”
“When I look at the way the tech world is going, I see a clear path to a future where AI is used to create something that is better than the best human beings,” it said.
Well, despite the uncomfortable warning earlier, it does seem that the AI acknowledges technology could bring about positive change as well. Maybe just like everything else, there are always positives and negatives to any new invention.
It’s time to breathe a sigh of relief that it doesn’t seem the robots will be taking over our planet just yet.
[via Mashable, cover image via Alexander Limbach | Dreamstime.com]
Also check out these recent news
