As Amazon, Google, Meta and Microsoft race to build AI tools; this 'biggest' AI question remains unanswer

There’s a race underway to build artificial general intelligence, a futuristic vision of machines that are as broadly smart as humans or at least can do many things as well as people can. However, as a report in news agency AP says, achieving such a concept — commonly referred to as AGI — is the driving mission of ChatGPT-maker OpenAI and a priority for research teams of tech giants Amazon, Google, Meta and Microsoft.This goal, often called AGI, is a driving force for these companies.However, progress in AGI has also raised concerns among world governments. A recent study published in Science by leading AI researchers warns that unchecked AI agents with advanced planning abilities could pose a serious threat to humanity.Defining the Elusive AGIBut what exactly is AGI, and how will we know when it's achieved? The term, once on the fringes of computer science, is now a popular buzzword with constantly evolving definitions.It's important to distinguish AGI from generative AI, which powers tools that produce new content like documents, images, and sounds. AGI is a broader concept.Geoffrey Hinton, a prominent AI scientist, emphasizes that AGI is not a technical term but "a serious, though ill-defined, concept." He defines it as AI that is at least as good as humans in most cognitive tasks.The Quest for a Thinking MachineThe term AGI originally aimed to revive the ambitions of mid-20th century computer scientists who envisioned intelligent machines. This was before AI research branched into specialized, commercially viable applications like facial recognition and voice assistants like Siri and Alexa."Mainstream AI research turned away from the original vision," says Pei Wang, a professor who teaches an AGI course. He highlights the desire of some researchers to "build a thinking machine" rather than just tools.Measuring Progress: A Work in ProgressThe lack of a clear definition makes it difficult to determine when or if AGI has been achieved.Hinton suggests that earlier, systems like GPT-4 or Google's Gemini might have been considered AGI for their ability to answer questions in a comprehensive way. However, as AI capabilities improve, the benchmarks for AGI are constantly being revised.The Road to AGI: More Than Just ChatbotsWhile impressive, today's chatbots powered by advanced autoregressive techniques are not quite the AGI envisioned by many. True AGI would require machines to excel in various tasks, including reasoning, planning, and learning from experiences.Researchers are exploring ways to assess AGI. A workshop dedicated to the topic is planned for a major AI conference next month. Organizers hope to establish a classification system for AGI progress, similar to how carmakers track the path towards self-driving vehicles.