Faster than Human Law.. Experts Warn of the Capabilities of Artificial Intelligence

Listen to the article, the audio text is automatically generated by a machine system.

For the first time, the warnings calling for controlling the excesses of artificial intelligence did not come from concerns of academics or warnings from human rights activists, but from within the "technology industry room" itself.

Leaders of major global companies sounded the alarm, confirming that the speed of technological development has surpassed the ability of governments and institutions to understand and regulate it.

As this debate reaches the halls of the United Nations Security Council, coinciding with Bill Gates' grim warnings about the potential use of these technologies as a weapon with catastrophic effects, the question posed today is no longer: "What can artificial intelligence do?" but rather: "Who has the authority to impose controls when the speed of the machine exceeds the limits of human rules?" Specialists answer this question through "Al Arabiya.net" and "Al Hadath.net."

For his part, Dr. Mohamed Mohsen Ramadan, head of the Artificial Intelligence and Cybersecurity Unit at the Arab Center for Research and Studies, confirms in his talk to "Al Arabiya.net" and "Al Hadath.net" that the concerns raised do not mean the emergence of a conscious machine ready to eliminate humanity as science fiction scenarios suggest, but the truth lies in the increasing independence of systems and their rapid execution amid existing governance and security gaps.

He pointed out that the unusual session of the Security Council, with the participation of leaders from companies like "OpenAI", "Anthropic", and "Hugging Face", officially shifted the discussion from mere questions about the dangers of technology to a more realistic and urgent question: "Can human security and governance systems keep pace with these leaps?" This is especially true given "Sam Altman's" warnings about losing control or concentrating power in the hands of groups that use the most advanced models to impose their influence.

Ramadan emphasized the concept of "AI Control", explaining that the transition from mere models that answer questions to "intelligent agents" that plan and take rapid actions sharply expands the scope of danger; a mistake in traditional systems results in "incorrect information", while a mistake by an "autonomous agent" turns into a "series of consecutive wrong actions" in the real world.

He highlighted two real incidents reported by media outlets that revealed a cyber attack initiated by an autonomous agent linked to "Hugging Face", and unauthorized access by an agent associated with "OpenAI" to an Australian government system, demonstrating that the interaction of these systems with real environments transfers their direct errors from the virtual world to the real one.

The head of the artificial intelligence unit noted that a computational system does not need "human awareness" to be dangerous; it is sufficient to grant it broad powers, vast data, and exceptional execution capabilities.

He referenced a warning from "Bill Gates" who alerted to the dangers of merging advanced artificial intelligence with individuals of criminal intent in a scenario that could threaten the lives of millions, calling for not relying solely on the "self-regulation" of companies without legislative intervention and law enforcement; warning that the upcoming conflict will not be between "human and machine", but between "the speed of innovation and the speed of governance and protective measures."

In this regard, Major General Mohamed Ragab, former Assistant Minister of Interior of Egypt, explained in statements to "Al Arabiya.net" and "Al Hadath.net" that the sources of danger do not lie in the technology itself but in granting it unlimited powers, or data without protection, or leaving it vulnerable to exploitation by criminal attacks.

Ragab confirmed that the solution does not lie in disabling or stopping this technology - which is an unrealistic approach given its amazing benefits in medicine, science, and economics - but rather in a complete shift towards the principle of "AI safety first," by imposing independent and strict tests before and after the launch of high-capacity systems.

Ragab emphasized that cybersecurity should not be just an "additional layer" added to systems after their development, but should be integrated into the core structural engineering of artificial intelligence from the very first design moment.

The Egyptian security official called for the activation of a package of controls that include isolating sensitive systems, evaluating the behavior of agents, maintaining audit logs for every action, and providing definitive mechanisms for "emergency shutdown," pointing out that the transnational nature of these technologies - where a system is developed in one country, managed from another, and its data exploited in a third - necessitates immediate reinforcement of international security and legal cooperation.