AI
DeepMind researcher warns AI may soon surpass human intelligence
A Google DeepMind researcher states that Artificial Superintelligence could emerge shortly after AGI, urging immediate global preparation for its societal impact.

Researchers at Google DeepMind have issued a warning that artificial intelligence is approaching a critical threshold, advocating for proactive consideration of the potential consequences. Tim Genewein, a researcher at the lab, notes that while the scientific community has long focused on Artificial General Intelligence (AGI)—machines capable of matching humans in most cognitive tasks—this milestone is becoming increasingly attainable.
The path to superintelligence
Genewein argues that AGI will likely be followed rapidly by Artificial Superintelligence (ASI). This stage implies that machines could outperform large numbers of humans across a wide spectrum of activities, ranging from accelerating scientific progress in entire fields to achieving significant practical feats such as designing electronic chips. The researcher emphasizes that reaching ASI does not merely mean superiority in specific tasks, but rather dominance in most areas relevant to humanity.
Despite this outlook, some experts remain skeptical about whether AGI can be achieved at all. They contend that current AI systems rely heavily on language, which limits their understanding of the real world. However, Genewein posits that "cognitive manufacturing" will drive the creation of far more powerful systems, fueled by improvements in chip design, increased software efficiency, and massive investments in computing infrastructure.
Computing power and key developments
According to this perspective, the computational capacity available to AI could increase tenfold annually, representing a substantial leap over a single decade. While such advancements may bring AI closer to human levels, they do not guarantee the overcoming of barriers that might prevent the arrival of ASI. Genewein illustrates this uncertainty by comparing Einstein’s discovery of general relativity with the inability of many ancient scientists to achieve similar breakthroughs.
The researcher identifies several developments that could alter the trajectory of AI evolution:
- The creation of new training data that simulates the real world.
- Technologies enabling systems to generate novel ideas.
- Enhanced capabilities for AI agents to collaborate.
- Iterative self-learning, allowing systems to train themselves.
Global coordination required
Genewein suggests these factors could gradually transition current intelligence to expert-level AGI, and subsequently to superintelligence. This progression would accelerate scientific advancement and open avenues for breakthroughs in medicine, materials science, and clean energy. He stresses the necessity for the world to begin considering how to reach ASI and what might follow, through broad, joint efforts involving advanced laboratories, national institutions, and international bodies.
The researcher concludes that human-level general intelligence is no longer the maximum possible limit. Understanding what comes next, along with its implications for science, industry, and society, represents the major upcoming challenge. Precise tools for measuring the speed of this progress are also deemed essential.
Latest news

Passing of Sheikh Abu Saleh Muhammad Al-Andari, Spiritual Leader of the Druze Community

Lebanon, Syria, Jordan and Iraq: Hani Seeks to Build Arab Agricultural Bloc Enhancing Food Integration

Salam: Independence Is Not Complete Without Restoring All Our Land!


