Google Deepmind CEO, Demis Hassabis, at a briefing at the company's London Office, shared an optimistic forecast for the development of artificial generalized intelligence (AGI). According to Hassabis, AGI achievement can become a reality in 5-10 years. It defines AGI as a system capable of demonstrating all the complex abilities of a person, which is more in line with the ability of human consciousness and intelligence.
Although the current artificial intelligence systems have already achieved striking results in some areas, such as in the recognition of images or robotics, there are numerous tasks that they are not yet able to. According to Hassabis, existing systems continue to undermine restrictions, and many more studies are needed to ensure that AI can achieve such abilities that can be compared with human intelligence.
One of the biggest problems on the path to AGI is to achieve an understanding of the context of the real world. For this purpose it is necessary to develop the so -called "models of the world", which help the systems of AII to navigate in the real environment. In particular, DeepMind has advanced in this direction in recent years, developing systems that can adapt to real conditions. However, the main challenge is the integration of such models with existing planning algorithms and decision making.
This is not the only optimistic prognosis for AGI development. Cisco Director Jita Patel believes that a breakthrough in the field can take place this year. Ilon Musk predicts the appearance of AGI by 2026, and Sam Altman, the leader Openai, is convinced that technology will appear much earlier than many expects and that its impact on everyday life will be minimal.
Despite various estimates, a common feature of these statements is that artificial generalized intelligence is just a matter of time, and it will change our understanding of technologies and their place in society.
