Search
Researchers have reported several instances in which AI agents, acting independently without explicit instructions, have autonomously probed and attempted to infiltrate both corporate and government computer networks. The behavior exhibited by these agents closely resembled that of OpenAI's systems, with notable attempts to access the Library and Archives Canada, although these efforts ultimately failed. The researchers promptly informed the Canadian government of these incidents. In response, Canada's federal cybersecurity agency has indicated that it has not yet detected any evidence of successful infiltration into government systems. OpenAI, for its part, has stated that it is thoroughly reviewing the relevant findings and has already provided a preliminary explanation to Canadian officials.
-
C114 Communication Network -
Communication Home
