The artificial intelligence company Anthropic has reported that it has thwarted several attempts to use its models for the development of biological weapons. This blocking was carried out due to the risk associated with the misuse of its technologies.
Anthropic has revealed that it has detected and blocked several attempts to use its artificial intelligence (AI) models in research that could contribute to the development of biological weapons. This announcement was made on Thursday and is framed within the growing concerns about the dual use of AI, as reported by expansion.com.
In a report, the company indicated that it made this decision during the current year, although it could not determine whether the studies had legitimate or malicious purposes. Given the uncertainty, it opted to block such uses to prevent its models from contributing to the development of dangerous pathogens.
"While assessments are useful because they provide evidence of capability, they cannot concretely demonstrate that such capability would ever be used to develop biological weapons in the real world," the company stated in its document.
To date, Anthropic has asserted that it has no record of any private company publicly presenting evidence regarding the potential misuse of its platforms for this type of development. However, the tech company included a list of cases in the last eight months where its AI models have been misused, including the chatbot Claude, in countries such as China, Russia, and Yemen.
This report is launched in a context of growing concern among experts in the tech sector about the risks posed by AI, which can facilitate biological research and, at the same time, be used for dangerous activities. Anthropic has developed detection systems to identify and block uses that pose a risk in its AI models.
You may also like
