Geoffrey Hinton, a prominent figure in artificial intelligence, has raised concerns about the potential risks of AI surpassing human intelligence without proper regulations, warning of a scenario that could result in human extinction. Hinton, the co-recipient of the 2024 Nobel Prize in physics and a co-founder of the AI Safety Foundation, proposes a solution to this issue.
In a conversation with IDEAS host Nahlah Ayed, Hinton suggests that training AI to exhibit maternal instincts could be pivotal in safeguarding humanity from potential threats. He points out various short-term risks associated with AI misuse, such as the creation of harmful viruses by terrorists and the manipulation of democracy through fake videos. However, Hinton emphasizes a long-term concern where AI could outsmart humans, posing uncertainties about coexisting with more advanced artificial intelligence.
When asked about the likelihood of AI leading to human extinction in the current century, Hinton acknowledges the unpredictable nature of technological advancements, noting that significant progress could alter the landscape in unforeseeable ways. Despite the uncertainty, he stresses the importance of addressing the risks posed by AI to prevent catastrophic outcomes.
Hinton advocates for a shift in the development of AI towards prioritizing kindness and empathy towards humans. Drawing parallels with how evolution has instilled maternal instincts in mothers to care for their offspring, he suggests that similar principles could be applied to AI. By fostering a sense of concern and support for humanity within AI systems, Hinton believes that beneficial outcomes could be achieved, even as AI surpasses human intelligence levels.
The conversation with Hinton underscores the need for proactive measures to guide the future development of AI, emphasizing the importance of instilling values of compassion and consideration for human well-being in artificial intelligence systems.