Geoffrey Hinton, the “godfather of AI” and former vice president of engineering at Google, believes that the world needs to find a way to control AI as it develops. In a recent interview, he acknowledged the fantastic potential of AI technology but stressed the importance of understanding how to contain it and avoid its negative consequences. Hinton also believes that AI poses several dangers such as job reduction, the creation of fake news, and the possibility that it could eventually become smarter than humans.
Hinton recommends putting as much effort into developing AI technology as we do into making sure it’s safe, something he says is not happening right now. He worries about the alignment problem – ensuring that AI has goals that benefit humans. While synthetic intelligence hasn’t evolved with innate goals, Hinton points to the potential for machines to generate their own goals, which could eventually lead to AI becoming more powerful than humans.
Hinton believes that the best approach to AI development is to have many intelligent minds work together to figure out how to contain its dangers. Although a recent letter calling for a six-month-long moratorium on training AI systems more powerful than OpenAI’s GPT-4 is “completely naive” according to Hinton, he emphasizes the need to recognize the potential dangers of AI and work to contain them.
Indeed, Hinton stresses the need to control AI as it advances, pointing out that there’s no use waiting for AI to outsmart us. Instead, we need creative and intelligent people to ensure that AI is kept in check before it gets too smart. Governments must insist on flagging fake images, and companies must act responsibly, even in a competitive system.
In terms of AI’s potential to surpass human intelligence, Hinton states it could take anywhere from 5 to 20 years, but he’s not confident about the prediction. Overall, Hinton stresses the importance of thinking about the implications of AI development seriously, emphasizing that we must find a way to control AI before it’s too late.
Editor Notes:
Geoffrey Hinton’s perspective on AI development is crucial as we continue to advance in this technology. While AI can help promote many advances, it’s vital that we be aware of the potential issues it poses and address them accordingly. As we continue to develop AI, we must create goals that benefit us, and we must work together to ensure AI does not become more powerful than humans. It’s exciting to see many brilliant minds working together to solve these problems, as this will lead to more beneficial AI systems. Visit GPT Newsroom to learn more about AI and its impact on our world.
Source link
from GPT News Room https://ift.tt/fu03BT8
No comments:
Post a Comment