Updated

US technology company Anthropic has raised concerns that smaller AI models may soon be capable of self-replication without requiring access to centralized data centres, potentially within a matter of months. The warning highlights emerging risks in artificial intelligence development as models become more distributed and autonomous.