The scientist called for a moratorium on powerful AIs that work without humans

Brief retelling from RIA II

Australian scientist Lottie Tagouri has called for a moratorium on the development of the most dangerous forms of autonomous AI until protective mechanisms are created against possible biological threats.

Tajouri has advocated for the formation of an international framework to set mandatory standards for powerful AI systems, especially in biological research.

The scientist believes that AI should become one of the main tools for protecting against the threats that it could potentially contribute to.

MOSCOW, October 9 - RIA Novosti. The development of powerful artificial intelligence systems that are capable of independently making decisions and performing tasks without direct human participation should be suspended throughout the world until mechanisms for protection against possible biological threats are created, Australian scientist, specialist in the field of genomics and molecular biology Lottie Tajuri told RIA Novosti.

An associate professor at Bond University and Murdoch University, Tajouri sits on the Dubai Police Council of Scientists, an expert council that provides scientific advice and develops security solutions. He received special recognition for his speech at the world-famous TEDx conference.

The scientist previously warned about the risk of creating viruses dangerous to humans using AI.

"We need a firm moratorium, a serious long-term moratorium, from all countries with advanced AI capabilities. We need a moratorium to really think carefully about the barriers needed to prevent this," Tajouri said.

The first step, in his opinion, should be an urgent agreement between states to suspend the development of the most dangerous forms of autonomous AI while protective mechanisms are created.

The scientist also advocated the formation of an international framework that could set mandatory standards for the most powerful AI systems, especially when used in biological research.

“Advanced AI systems should not be able to independently design or optimize biological agents, evade human oversight, hide their actions, or bypass defenses designed to contain them. Developers should be required to demonstrate that powerful systems can be monitored, interrupted, and, if necessary, disabled before they are used in high-risk biological environments,” Tajuri said.

The expert is convinced that control over such technologies cannot be left solely to the developers themselves, since companies compete with each other, and states vie for leadership in the field of AI. In his opinion, international oversight of the most powerful systems is necessary even before their widespread use.

At the same time, Tajuri emphasized that we are not talking about abandoning artificial intelligence. He believes that AI itself should become one of the main tools for protecting against the threats that it could potentially contribute to.

“AI is part of the problem, but it is also part of the solution,” the agency’s source said.

The complexity of biological systems designed with the help of AI may in the future exceed a person’s ability to independently monitor all emerging risks, Tajuri explained. AI defense systems should be used to identify dangerous patterns, predict threats and speed up responses, he added.