Dario Amodei, CEO of Anthropic, has warned in his latest statements about the dangers arising from the development of superintelligent AI and has urged related industries to slow down the pace of this technology's development. This issue became a serious discussion after one of the company's researchers reported serious dangers arising from competition in this field.
Concerns About Self-Improving AI
The recent warnings were particularly raised after one of Anthropic's researchers stated that leading companies in the field of AI are "running towards superintelligent self-improvement and gambling with our lives." He also estimated that the likelihood of this technology leading to the extinction of humanity in the next decade is over 10 percent.
Read more: Strategic Cooperation Between Oman and the Islamic Coalition Against Terrorism
In this context, Amodei called for serious attention to the pace of AI development. Other senior executives, including Sam Altman, CEO of OpenAI, and Elon Musk, CEO of xAI, also supported this request. In contrast, Donald Trump, the former President of the United States, opposed slowing down AI development in the U.S., emphasizing competition with China.
Potential Consequences of Superintelligent AI
The hypothetical technology of superintelligent AI generally refers to a system that possesses intelligence superior to the best human minds. This technology could soon independently engage in scientific research, manage companies, and design new technologies. The existing concerns are more related to the pathway that some researchers believe leads to this level of AI: automatic self-improvement where AI systems are used to design and train their next generation.
The International AI Safety Report in 2026, led by Yoshua Bengio, a Turing Award winner, has stated that current AI systems still lack the necessary capabilities for loss of control scenarios but are rapidly advancing.
Among the potential consequences of this advancement are the lack of technical control and the surge of biological weapons and cyberattacks. Advanced models could be used by hostile actors to engineer pathogens or carry out large-scale cyberattacks against critical infrastructure.
Moreover, these advancements could lead to geopolitical and economic instabilities. A sudden leap towards superintelligent AI could disrupt labor markets and concentrate geopolitical power in the hands of a few companies or countries.
At the same time, these technologies could also bring significant benefits. Proponents of superintelligent AI believe that these technologies could be effective in treating diseases, accelerating scientific discoveries, and helping to solve global problems such as climate change.
Ultimately, Amodei, in an article titled "We Must Set Boundaries," has called for leading companies to provide greater access to independent security assessments. This request has become a rare agreement among technology executives, with Altman and Musk also supporting it.
Read more: The Shanghai Cooperation Organization's Anti-Terrorism Group Condemns Recent U.S. Attacks on Iran · Satya Nadella: Human Control Over AI is Essential
