Tim Genewein, a researcher at Google DeepMind, has raised an alert regarding the rapid advancements in artificial intelligence (AI). According to executivedigest.sapo.pt, Genewein suggests that AI is on the brink of surpassing human cognitive abilities, marking a significant milestone in technological research. This development could lead to the emergence of superintelligent machines capable of outperforming humans in a wide range of intellectual tasks.
Genewein points out that the transition from general AI, which matches human performance in cognitive tasks, to superintelligence, where machines exceed human capabilities, might happen in a relatively short timeframe. He describes this process as 'cognitive industrialization', driven by three key factors: advancements in semiconductor technology, increased efficiency of AI systems, and substantial investments in tech infrastructure. These factors are contributing to a tenfold increase in AI processing capacity each year.
Despite this rapid growth, Genewein emphasizes that more computational resources do not automatically result in more intelligent systems. He uses the example of Albert Einstein to illustrate that sheer computational power does not guarantee innovative discoveries. However, developments such as new training data types, technologies for generating original ideas, and AI agents capable of effective collaboration could help overcome these limitations. One transformative possibility is recursive self-improvement, where AI systems enhance their own training and development, potentially accelerating technological progress.
Written by the devs.com.pt editors from the sources below.