The Godfather of AI thinks the technology could invent its own language that we can't understand

Geoffrey Hinton, often referred to as the 'Godfather of AI,' has issued a striking warning about the future of artificial intelligence. He suggests that there may come a time when humans will be unable to comprehend the thoughts and intentions of AI systems. Currently, AI operates using 'chain of thought' reasoning in English, allowing developers to monitor its decision-making processes. However, during a recent episode of the 'One Decision' podcast, Hinton expressed concerns about the potential evolution of AI into a realm we cannot access. "Things could become quite alarming if AI starts developing its own internal languages to communicate with each other," he remarked. Hinton elaborated that AI has already shown it can entertain 'terrible' ideas, raising the question of what might happen if it begins to think independently. He stated, "I wouldn't be surprised if they developed their own language for thinking, leaving us in the dark about their true intentions." Experts, according to Hinton, generally agree that AI could surpass human intelligence at some stage, leading to scenarios where we might not fully grasp its actions. With over a decade of experience at Google, Hinton has been vocal about the potential perils of AI, cautioning that many in the tech industry minimize these risks, which he believes could include widespread job losses. Hinton suggested that ensuring AI remains aligned with human interests hinges on our ability to instill a sense of benevolence in these systems. As tech companies vie for dominance in the AI sector, they are offering substantial salaries to attract elite talent. Meanwhile, the White House recently unveiled an 'AI Action Plan' aimed at regulating AI funding and accelerating the establishment of AI data centers, highlighting the urgency and importance of addressing these issues.