Google AI CEO Demis Hassabis has given a big warning regarding the future of Artificial Intelligence. He says that the world has come very close to Artificial General Intelligence i.e. AGI and we have only a few years left to prepare for it. According to Hassabis, AGI could become a reality within the next four years. He has appealed to policy makers, economists and common people to be ready for this change from now on.
Google AI CEO’s big warning regarding AGI
Demis Hassabis said in a report that humanity is entering a period of technological change where the arrival of AGI is not far off. AGI means AI systems that can achieve human-like understanding and capabilities in many cases. According to Hassabis, the world is currently standing in the initial stages of “Singularity” and the next few years could prove to be very important. He predicted that AGI could emerge within the next four years. This is why he has advised governments, industries and common people to be mentally and technically prepared for this change. He believes that in the coming years, the impact of AI will be visible on almost every sector of society and economy.
Upcoming AI agents will be a big test
Hassabis said that the AI agents seen in the next year should be seen as preparation for even more powerful systems of the future. According to him, this round is a kind of exercise, through which the society will be able to understand how to keep pace with advanced AI technology. He also said that the AI systems that people consider to be very powerful today may be just the beginning signs of the technologies to come in the future. Hassabis believes that current AI models are a glimpse of major changes to come to society. Therefore, it is important that along with technological development, its social impacts should also be seriously discussed.
Security, regulations and future challenges
The Google AI chief admitted that he deliberately used strong and warning language so that policymakers and the public would take the topic seriously. He welcomed the initial initiatives of the US government which call for security testing of new AI models before making them available to the general public. Hassabis also said that completely self-correcting AI systems do not yet exist, but this possibility is not decades away. According to him, in the near future we may see such AI models which can improve their capabilities on their own. In such a situation, it will be very important to give priority to security, surveillance and responsible development so that AI can be used in the interest of humanity and possible risks can be controlled in time.
Also read: 32 inch smart TV launched for just Rs 11,499, equipped with 48W sound and Google TV 5.0
Leave a Reply