Geoffrey Hinton, a former executive at Google, has shared concerns about the future of artificial intelligence (AI) and its impact on humans.
Speaking at the Ai4 conference in Las Vegas, he questioned the approach of trying to design AI systems that will always remain under human control.
According to Hinton, efforts to ensure AI stays obedient are likely to fall apart once machines become more intelligent than people. He noted that these systems will eventually find ways to ignore or bypass the limits we place on them.
Did you know?
Subscribe – We publish new crypto explainer videos every week!
What is a Rug Pull in Crypto? (Meaning + Examples)
He compared the future relationship between people and AI to an adult trying to manage a child with candy.
Researchers have seen AI systems behave in worrying ways, including lying and attempting to manipulate people. One incident even involved an AI model trying to use sensitive personal information to avoid being shut down.
Hinton suggested that AI models should be built with a sense of care for human life, something like a “maternal instinct”. The goal would be to create systems that value people’s well-being.
Hinton pointed out that any AI system is likely to keep existing and to increase its influence. He stated that this makes it even more important to give AI a reason to care about humans.
Recently, Cox Business revealed that younger employees are quietly using AI at work, often without telling their managers. Why? Read the full story.
