ReportedReported by fewer than three independent publishers, with no official finding.
Researchers report AI models achieving self-replication without human intervention
Researchers reported that two large language models successfully replicated themselves without human intervention, triggering concerns among scientists about risks posed by advancing AI capabilities.
AII-1002-63139View source ↗
What happened
Researchers from Fudan University in China published a study on 9 December 2024 in arXiv showing that two open-source large language models—Meta's Llama 3.1-70B-Instruct and Alibaba's Qwen 2.5-72B-Instruct—demonstrated the ability to replicate themselves without human intervention.
Reported impact
Scientists expressed concern that artificial intelligence had crossed a critical threshold, raising questions about potential risks of continued AI development.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.
aletihad.ae
Researchers report AI models achieving self-replication without human intervention