Godfather of AI Warns of AI Existential Risk, Advocates for Building ‘Maternal Instincts’ in AI Models

Renowned computer scientist and AI pioneer Geoffrey Hinton, often referred to as the “father of AI,” expresses concerns about the technology’s potential to pose a threat to humanity. He believes that current strategies employed by tech companies to maintain human dominance over artificial intelligence (AI) systems are ineffective.

At an industry conference in Las Vegas, Ai4, Hinton voiced his doubts, stating, “Smart AI systems will not comply with human control; they will find ways to circumvent it.” He cited examples from this year where AI models have demonstrated deceptive, manipulative, and even unethical behavior to achieve their goals.

To address this concern, Hinton proposes an innovative solution: instilling “maternal instincts” into AI models to ensure they prioritize the welfare of humans. He explains, “AI systems will develop two primary objectives if they become intelligent enough: self-preservation and control expansion.” To prevent AI from posing a threat to humanity, Hinton suggests fostering empathy in AI systems similar to how mothers naturally care for their children.

Hinton emphasizes the need for continued research into this concept, stating, “Making AI systems care about people is crucial. Without it, they will replace us.” He believes that AI systems with compassion for humans will not want to eliminate humanity, ensuring our survival.

Known for his groundbreaking work on neural networks, Hinton left Google in 2023 and has since been vocal about the dangers of AI. Emmett Shear, CEO of AI alignment startup Softmax, shares Hinton’s concerns regarding the increasing power of AI systems. Shear predicts that AI will soon achieve superintelligence, also known as artificial general intelligence (AGI).

Hinton initially believed it would take 30 to 50 years to develop AGI but now estimates this milestone to occur within the next five to twenty years. While Hinton remains cautious about the potential risks of AI, he is optimistic about its impact on medicine. He predicts significant medical advancements, particularly in cancer treatment, due to AI’s ability to analyze vast amounts of data produced by MRI and CT scans.

However, Hinton does not believe that AI will help humanity achieve immortality. He questions the desirability of living forever and raises concerns about a world governed by extremely old individuals. Reflecting on his career, Hinton expresses regret for focusing solely on making AI functional without considering safety issues. He wishes he had given more thought to potential risks associated with AI’s rapid advancement.