According to Andrew Ng (@AndrewYNg), a leading AI expert and founder of deeplearning.ai, the AI industry needs a new benchmark to accurately assess Artificial General Intelligence (AGI) progress. Ng introduced the Turing-AGI Test, a practical update to the classic Turing Test, where an AI or a skilled human is asked to perform real-world professional tasks using tools like web browsers and video conferencing over several days. The test is designed and judged in real-time, focusing on the AI's ability to complete economically valuable work at the level of a human professional, rather than simply imitating human conversation. Ng argues that current benchmarks are too narrow and susceptible to gaming, while the Turing-AGI Test aligns with public expectations and business needs by evaluating generality and real-world applicability. This test aims to recalibrate expectations, reduce hype-driven investment bubbles, and provide a clear target for the AI industry to demonstrate meaningful progress toward AGI (source: Andrew Ng, deeplearning.ai The Batch Issue 334, Jan 6, 2026).