AI’s rapid advances offer numerous development opportunities, but also raise the question of how planning and the refinement of institutions and policies can keep pace with technological progress, at a time when high expectations often come with high risks.
After a period in which AI technology took the world by storm amid an intense race to develop AI models, a paradox is now emerging in the field: leaders of major technology corporations, many of which once took a determined part in the race, are calling for a slowdown in AI development.
This paradox reflects an important shift, as corporations gradually move their focus from a race to advance technology to a race to control the pace of development and the risks arising from the technology.
In a recent article, Anthropic CEO Dario Amodei called on leading AI companies to slow the pace at which they enhance the capabilities of AI models, arguing that companies need to allow more time for safeguards to keep pace with the growing capabilities of the models. The call has received support from a number of AI companies, including OpenAI.
Current developments show that AI is no longer simply a technological issue but is becoming closely linked to global security. The phrase “out of control” is increasingly used to describe the potential risks posed by the technology. Numerous reports indicate that AI models are being used for dangerous purposes.
According to the Anthropic executive, within the next six to 12 months, a large group of AI agents could potentially become capable of coordinating to take control of online systems through a botnet – a network of electronic devices such as computers and phones that have been infected with malicious code and are remotely controlled by malicious actors without the owners’ knowledge. This could cause enormous economic damage.
The Secretariat of the UN Security Council Counter-Terrorism Committee has warned that terrorist organisations are using digital technologies, including AI, to support recruitment activities, particularly the recruitment of children and young people.
The enormous benefits brought by AI are undeniable, but it is necessary to ensure that people remain at the centre of the governance of this technology. Scientists have stressed the urgent need to tighten oversight and establish mechanisms to ensure the safety of the technology.
The role of governments in regulating AI has long been debated. However, in the current discussion about controlling the pace of AI development, the role of technology companies is receiving considerable attention.
No company wants to fall behind in developing advanced AI models, which are regarded as key products and cash cows for many businesses. In his essay, Amodei therefore proposed that AI developers work together to establish common standards on the pace of model development, preventing competition between companies from becoming a driving force that causes safety measures to be overlooked. He also acknowledged that adjusting the pace of AI development would not be easy.
Global investment in AI is forecast to exceed 1 trillion USD in 2026, bringing high expectations for breakthroughs and major advances across various aspects of life.
At the same time, however, discussions are focusing not only on growth potential but increasingly on controlling technological risks, including the development of AI governance frameworks covering responsibility and ethics, risk and safety, and international cooperation, among others.
To ensure a safe digital environment for everyone, a comprehensive strategy and the participation and contribution of all stakeholders are needed.
