Artificial intelligence research company Ntropic has revealed that scientists have tried to use AI models to research possible biological weapons.

The company says that after this effort by scientists, the company has made its security systems more stringent.

The company said in a report that this year it halted several potential projects in which customers were trying to use the company’s AI models for advanced biological research.

According to Ann Tropic, she could not determine whether the research was actually being conducted to develop biological weapons.

This distinction is particularly difficult in the field of biology, because the same type of research can have both medical and legitimate purposes, such as the development of a vaccine, while in some cases the same research may also contribute to making the germ more dangerous.

The company said that is why they decided to take the strictest measures while taking precautionary measures.

The researchers were accused of trying to circumvent the security system

Researchers involved in these incidents related to biological research took steps to hide their activities and avoid restrictions on AI systems, N. Tropik said.

According to the company, the scientists tried to obfuscate the purpose of their research to avoid our security systems.

NThropik subsequently suspended accounts associated with these activities.

The company said that experiments conducted last year with older versions of its AI models showed that they were not capable of providing meaningful support in risky biological research, however, according to the company, the new systems are capable of performing far more complex scientific tasks.

In view of this increase in capacity, the company has imposed stricter restrictions on biological queries that may potentially have dual-use purposes.

Weapons manufacturing and propaganda are also among other misuses of AI

The biological research incidents were part of a broader report by Ntropic, which detailed various examples of misuse of AI systems, including Cloud, over an eight-month period.

The company said it also discovered attempts to use the cloud to develop conventional weapons, including software related to firearms, missiles, armed drones and bombs.

According to the company, six such incidents were linked to users in China, Russia and Yemen.