A notable shift is underway in artificial intelligence: some of the field's leading builders are turning their attention away from text-based chatbots and toward "physical AI" — software that powers robots and other machines operating in the real world.
According to ABC News, top developers are pivoting from chatbots to physical AI, signaling that the momentum in the industry may be moving from screens to machines that can sense and act in physical space.
The trend is especially visible in South Korea, where the push into physical AI is raising pointed questions for the people building these systems. According to KoreaTechDesk, Korean founders working on physical AI are now facing a hard question about benchmark scores — the standardized tests used to measure how well an AI system performs.
That tension matters because benchmarks have long been the currency of credibility in AI. Impressive scores help startups attract funding, customers, and talent. But when AI moves out of the chat window and into robots that have to function reliably in messy, unpredictable physical environments, it becomes less clear whether high benchmark numbers actually translate into machines that work well in practice. Founders, according to KoreaTechDesk, are being pressed to reckon with that gap.
Together, the two reports sketch a picture of an industry in transition. The race that defined recent years — building ever-more-capable conversational systems — is giving way to a harder, more physical challenge: getting AI to move, manipulate, and operate in the real world.
Why it matters: If the next wave of AI is about robots rather than chatbots, the way we measure success may need to change too — and that shift could shape which companies, and which countries, lead the field.