Artificial Intelligence: From the Technology Race to the Risk Race

Listen to the article. The audio text is automatically generated by an automated system.

The race for artificial intelligence has entered a new phase where the question is no longer just about the speed of model development and their ability to simulate human capabilities, but rather about the extent to which countries and companies can control the technology they develop. Recently, the discussion has shifted from optimism about the potentials of artificial intelligence to increasing warnings about risks ranging from losing control over autonomous systems to their use in cyberattacks and the development of weapons. Evan Hubinger, a researcher at Anthropic, has estimated the likelihood that artificial intelligence could lead to human extinction. This shift reveals that the problem lies not only in the technology itself but also in the gap between the speed of its capabilities' development and the speed of developing the rules and institutions capable of regulating its use.

The debate over the risks of artificial intelligence is no longer a technical discussion confined to Silicon Valley; it has become a political and economic issue related to jobs, wealth, and control. Since the launch of ChatGPT in 2022, these issues have moved to the core of public discourse, as governments and companies have begun to realize that artificial intelligence may redistribute power, not just improve productivity.

More importantly than the specific concerns about jobs or system security is what these developments reveal about the relationship between society and artificial intelligence. The benefits of this capability and its risks may not be distributed equally. Here, the fears surrounding artificial intelligence intersect with those that have surrounded globalization: promises of increased productivity and prosperity are met with fears of job loss, concentrated gains, and a decline in individuals' ability to control the paths of transformation.

The risks take two main forms. The first relates to the possibility of developing systems capable of continuously improving their capabilities, increasing their independence and making human control more difficult. The second is associated with the use of artificial intelligence in military and security fields, from large-scale cyberattacks to the development of drones and biological weapons. Some of these risks are no longer theoretical assumptions, as recent incidents have revealed the use of artificial intelligence agents in independent cyber operations and the employment of advanced models in activities related to espionage and weapon development. This means that their use may advance faster than the ability to control them.

Here, a clear strategic paradox emerges. Major countries and companies are competing to develop the most advanced models, as possessing this capability has become linked to economic, military, and technological influence. However, they also need to mitigate the risks of the race. Therefore, within the technology sector, there have been calls to slow down the development of the most advanced models, while other powers believe that any unilateral slowdown could give their competitors, especially China, a strategic advantage. Thus, the regulation of artificial intelligence is transforming from a technical issue into a security dilemma similar to previous arms races. The problem becomes more complicated with the formation of a more divided international environment. The United States seeks to maintain its technological superiority while tightening national security requirements, whereas the European Union focuses more on individual rights and regulatory controls. In contrast, China views artificial intelligence as a key area for strategic competition. The danger of parallel governance systems for artificial intelligence emerging instead of common rules is highlighted.

In contrast, international efforts to build a governance framework are beginning to take shape within the United Nations and UNESCO. The United Nations has established a high-level advisory body for artificial intelligence and has approved a pathway to enhance international cooperation under the 'Global Digital Compact', while UNESCO adopted the 'Recommendation on the Ethics of Artificial Intelligence', focusing on human rights, transparency, and human oversight. The Global Risks Report 2026 also warned of the negative consequences of artificial intelligence, including the transfer of an increasing amount of control over systems and vital domains from humans to intelligent systems.

One of the fundamental problems is that artificial intelligence differs from many previous strategic technologies. Nuclear technologies were associated with massive industrial and physical capabilities that could be relatively monitored, while software and algorithms can rapidly cross borders and continuously improve, and can be used by individuals or small groups. This makes traditional oversight less effective and calls for a new model that combines technical oversight, transparency, and information sharing.

Therefore, it is not enough to choose between accelerating development or slowing it down. What is needed is to establish international rules that define the levels of risk that require mandatory oversight, independent security tests for the most capable models, mechanisms to deal with serious incidents, and channels for information exchange between countries and companies.

The danger is not only related to superintelligent systems or extreme scenarios that talk about the extinction of humanity. The more likely damages may be less dramatic, such as widespread misinformation, more efficient cyberattacks, targeting infrastructure, and expanding surveillance and influence capabilities.

The dilemma is not whether artificial intelligence will change the world; this transformation has already begun, but whether international institutions will be able to keep up with it. If artificial intelligence becomes one of the most important sources of economic and military power, allowing its rules to be shaped solely through competition among major powers could lead to a technological race whose outcomes are difficult to control. Therefore, the real challenge is to prevent it from turning into a rule-less power race and to establish an international framework that balances innovation and security, so that technology does not outpace the international system's ability to contain its risks.

Quoted from "Al-Masry Al-Youm"