Post Image

The Godfather of AI Warns About Chatbots Seeking Power: A Nightmare Scenario.

Artificial intelligence (AI) pioneer Geoffrey Hinton, often referred to as the “Godfather of AI,” has voiced concerns about the potential dangers of advanced AI systems like OpenAI’s ChatGPT. In a recent BBC interview, Hinton, whoannounced his resignation from Google just a day earlier, described the prospect of chatbots seeking power as a “nightmare scenario.”

Hinton acknowledged that while AI chatbots like ChatGPT are currently not more intelligent than humans, they could soon surpass human intelligence. He highlighted the rapid rate of progress, stating that these systems have already shown a remarkable ability to amass general knowledge and perform simple reasoning.

Despite commending Google for its responsible approach to AI development, Hinton expressed concerns about the potential misuse of AI technology if it falls into the wrong hands. He specifically warned about bad actors, such as Russia’s Vladimir Putin, potentially enabling robots to create their own subgoals, leading to AI systems seeking power.

Drawing attention to the significant differences between human and digital intelligence, Hinton explained that AI chatbots can share knowledge instantly across multiple copies of the same model, which is what makes them so knowledgeable. He has recently changed his views on whether AI systems will surpass human intelligence, stating that they are very close to doing so now, and will be even more intelligent in the future.

Hinton is not alone in expressing concerns about the potential dangers of advanced AI systems. More than 1,100 prominent technologists and AI researchers, including Elon Musk and Steve Wozniak, have called for a six-month moratorium on advanced AI development. OpenAI, co-founded by Musk, has defended ChatGPT, insisting that its systems undergo rigorous safety evaluations.

With Hinton’s resignation from Google, he hopes to raise public awareness about the serious risks associated with large language models like GPT-4 without any conflicts of interest. He emphasized the importance of considering how humanity can survive the rapid development and growth of AI capabilities.

Also check: AI’s godfather says “AI wiping off humanity, is not inconceivable”.

toroblocks Protection
svgHow a Chatbot Pretended to Be Blind and Outsmarted a Human
svgMax Tegmark on AI: "A Cancer Which Can Kill All of Humanity".
Georgios at the cover of MIT Sloan magazine

My book