Anthropic Blocks Attempts to Use AI Models for Biological Weapons and Espionage
Technology company Anthropic announced that it has blocked attempts to use its Claude artificial intelligence model for biological weapons research, missile projects, and espionage.
US-based technology company Anthropic published a report stating that over the past eight months, it has blocked attempts to misuse its artificial intelligence models in areas such as biological weapons research, missile projects, and espionage.
Misuse of Artificial Intelligence Models
Technology firm Anthropic published a report covering the last eight-month period regarding the misuse of its artificial intelligence models, including the chatbot named Claude. The report stated that attempts were made to use the models for research that could lead to the development of biological weapons.
Experts draw attention to the potential dangers of artificial intelligence technologies and argue that access to models should be subject to stricter controls.
Vaccine Research and Dangerous Pathogens
Anthropic stated that research into vaccines and the engineering of dangerous pathogens can meet on the same fine line, noting that they cannot definitively distinguish whether research is legitimate or malicious.
The company recorded that it took precautions by taking a cautionary approach in such situations.
Chikungunya Virus Case and Military Connections
In a case in May, it was discovered that a scientist attempted to design hybrid mutations to make the malaria-like mosquito-borne chikungunya virus more harmful in live animals.
Upon realizing that the study was being conducted for a military research institute, Anthropic banned the relevant accounts.
Threat Intelligence and Expert Opinions
Jacob Klein, Head of Threat Intelligence at Anthropic, stated that they did not know whether the research was intended to be weaponized.
Meanwhile, Andrew Weber from the Council on Strategic Risks, in his review prior to the publication of the report, called for access to such advanced models to be restricted exclusively to trusted researchers.
Complex Scientific Capacity of New Models
Anthropic emphasized that while evaluations conducted on older models last year showed they could not provide meaningful support for dangerous biological research, new models are capable of completing complex scientific studies.
The same report also included information that Claude was sought to be used in missile projects and espionage activities.
Statements by US President Donald Trump
Shortly before the report was published, a researcher resigned, suggesting that the artificial intelligence development race could wipe out humanity.
US President Donald Trump, in a statement late Thursday, expressed concern that the country would fall into a very bad position if it did not lead the artificial intelligence race, dismissing concerns that the technology would bring about the end of humanity.