00:00Anthropic says it blocked attempts to misuse its AI models for potentially dangerous biological
00:06research, including work involving viruses that could be made more harmful. The company says it
00:12detected activity from multiple types of bad actors seeking to use its technology for biological
00:18and cyber-related purposes. Anthropic says newer models have stronger safeguards designed to prevent
00:25assistance with dangerous dual-use research. The disclosure adds to growing concerns over how
00:31increasingly capable AI systems could be exploited for real-world harm.
Comments