Mr. Eric Schmidt, former CEO of Google - Photo: Shannon Stapleton/Reuters.
In a remarkable statement, former Google CEO Eric Schmidt said that artificial intelligence (AI) is entering a period of rapid development, to the point where it can soon learn and improve without human guidance.
Speaking at a conference of the strategic consulting organization Special Competitive Studies Project (a project he founded), the former CEO commented that current AI systems are starting to operate independently, can learn by themselves, improve by themselves and even plan proactively.
“Computers are now learning to improve themselves. They’re learning to plan, and they don’t need to listen to us anymore,” he stressed. Schmidt calls this phenomenon “recursive self-improvement,” a process in which AI formulates hypotheses, tests them in robotic labs, and continues to refine its research without any human intervention.
Schmidt, who served as Google's CEO from 2001 to 2011 and then as executive chairman until 2017, also predicted an upcoming seismic shift in the labor market.
According to him, within just a year, AI could replace most - if not all - of today's programmers. Not only that, AI is progressing so fast that it can surpass even the best experts in fields like mathematics.
“We believe that AI is undervalued, not overhyped as many people think,” Schmidt said, citing tools like ChatGPT, Claude, Gemini, and Deepseek—which, while not specifically trained for programming, are widely used for that purpose.
Once it can refine itself, AI will explode at an incredible rate, even beyond human control - Illustration.
Not only stopping at technology, Schmidt also warned the US House Committee on Energy and Commerce about the great challenge in the energy sector. He emphasized that at the current rate of development, the demand for electricity for AI will soon exceed the capacity of the current power grid system to meet.
“People are thinking about building data centers that consume up to 10 gigawatts. Meanwhile, an average US nuclear power plant only provides about 1 gigawatt,” he said.
Facing the risk of falling behind, he called on the US government to quickly adjust its energy policy, increasing investment in both renewable and non-renewable energy sources to ensure the country is not left behind by the major powers.
Schmidt also stressed the need for increased oversight of AI models, especially open-source platforms, which, he said, could pose a serious threat to national security if not properly regulated.
However, he also affirmed that AI still needs human governance and relies on high-quality data to maximize its potential.
“Scientists should be in control, and AI should only be a supporting tool — that's the right order,” Schmidt concluded.
According to VTV
Source: https://baothanhhoa.vn/former-chairman-of-google-canh-bao-ai-se-som-phot-lo-chi-dan-cua-con-nguoi-co-the-thay-the-lap-trinh-vien-trong-vong-1-nam-nua-246117.htm
Comment (0)