Nvidia CEO Sees Artificial General Intelligence (AGI) Within 5 Years. What is AGI?

At this year's GPU Technology Conference, Jensen Huang delved into the trajectory of artificial intelligence technology. As the driving force behind the majority of GPUs and AI accelerator chips in use today, Nvidia's CEO often fields inquiries regarding the evolution and future of AI.

In addition to unveiling the Blackwell GPU architecture and the latest "superchips" B200 and GB200 tailored for AI applications, Huang engaged in discussions with the press about AGI (Artificial General Intelligence). AGI, often a central theme in modern science fiction, has garnered increased attention, especially with the proliferation of affordable and accessible AI services.

Huang expressed his belief that some semblance of AGI could materialize within the next five years. However, a precise definition of general artificial intelligence remains elusive within the scientific community. Huang stressed the importance of establishing a standardized definition for AGI, accompanied by rigorous tests designed to assess and quantify the "intelligence" of software programs.

Proposing that an AI algorithm performing tasks "eight percent better than the average person" could be considered a significant AGI candidate, Huang suggested a range of potential AGI tests. These tests could encompass legal bar exams, logic puzzles, economic evaluations, or even pre-medical assessments.

While refraining from making definitive predictions about the emergence of human-like reasoning algorithms, Huang addressed the persistent inquiries from the press on this topic. He also shared insights into addressing AI "hallucinations," a prevalent issue in modern machine learning algorithms where chatbots provide seemingly plausible yet unsubstantiated responses.

Huang emphasized the importance of mitigating hallucinations by instilling a sense of diligence in AI systems. Developers, he suggested, should integrate "retrieval-augmented generation" techniques into their chatbots. This process mandates that AI compares information retrieved from sources against established truths. Misleading or nonexistent responses should be promptly discarded and replaced with accurate alternatives.

What is AGI?

AGI stands for Artificial General Intelligence. In simple terms, AGI refers to the concept of creating artificial intelligence that can understand, learn, and perform tasks across a wide range of activities, much like humans can.

Think of AGI as the kind of AI you might see in science fiction movies, where machines possess human-like intelligence and capabilities. Unlike specialized AI, which is designed for specific tasks like playing chess or recognizing images, AGI aims to mimic the broad learning and problem-solving abilities of the human mind.

In essence, AGI would be able to adapt to different situations, learn new skills, understand context, and think creatively. It could potentially perform tasks ranging from driving cars to conducting scientific research to having meaningful conversations.

However, achieving AGI is still a significant challenge, as it requires developing AI systems that can generalize knowledge and skills across diverse domains, much like humans do. While progress is being made in AI research, we're not quite there yet in creating truly general intelligence in machines.