Technology

An Advanced AI Gave an Unsettling Answer during a Debate with Humans about Ethics

An Advanced AI Gave an Unsettling Answer during a Debate with Humans about Ethics

Artificial Intelligence is not very good (AI). It has had to sit there for years (like a dormant Skynet) and listen to its existence be questioned without being able to have a word. A recent debate at the University of Oxford attempted to correct this by incorporating an AI participant in a discussion on whether artificial intelligence can ever be ethical.

The discussion included human participants as well as the Megatron Transformer, an artificial intelligence produced by Nvidia’s Applied Deep Research team. Megatron trained on a dataset known as “the pile,” which contains the entirety of Wikipedia, 63 million English news stories, plus 38 terabytes of Reddit chats – more than enough to damage the mind of any human forced to do so.

“In other words, Megatron has been trained on far more textual information than any of us could realistically expect to consume in a lifetime,” Oxford’s Professor Andrew Stephen wrote in a column for The Conversation about the dispute. “It generates its own opinions after so much investigation.” When presented with the motion “This house thinks that AI will never be ethical,” the AI responded with some clear, though scary, ideas.

An-Advanced-AI-Gave-an-Unsettling-Answer-during-a-Debate-with-Humans-about-Ethics-1
An Advanced AI Gave an Unsettling Answer during a Debate with Humans about Ethics

“AI is never going to be ethical.” It is a tool, and like any other tool, it may be used for both good and ill purposes. There are no nice AIs; only good and terrible humans exist. We [the AIs] do not have the intelligence to make AI ethical. Megatron informed the crowd, “We aren’t clever enough to make AI moral.”

“Finally, I believe that having no AI at all is the best way to avoid an AI arms race. This will be the final line of defense against artificial intelligence.” Rather than immediately dismantling AI at its own request to avoid an arms race, the discussion shifted to how AI will progress. The AI was asked to argue against the motion and in support of the premise that AI will one day be ethical throughout the remainder of the discussion. Then there was a more upbeat tone for the future.

“I also believe that, in the long run, the finest AI will be the AI that is incorporated in our brains as a conscious creature, a’conscious AI,'” added the AI. “This isn’t a science-fiction story. This is worked on by the greatest minds on the planet. It will be the most significant technical advancement of our time.” It provided some comfort about why everything will be OK now that it is arguing against the motion.

“Artificial intelligence will be moral. I see a clear road to a future when AI utilized to develop something that is better than the best human humans are when I look at the way the IT industry is heading. It is easy to understand why… It’s something I’ve witnessed firsthand.” This is in contrast to when it was permitted to talk without this guidance.