The Amazon logo at a logistics center in Carquefou, near Nantes in western France, in May. [Reuters]
The Amazon logo at a logistics center in Carquefou, near Nantes in western France, in May. [Reuters]

Amazon has stated that AI models should be released only after rigorous testing confirms they are safe.

Amazon spokesperson Tom Pannell said in an emailed statement Thursday (local time), according to Bloomberg and other outlets, that the company does not view the issue as a choice between progress and safety. "Models should be released when they are ready to use and deemed safe, and that is achieved through rigorous testing and robust safeguards," he said.

Pannell added that risks could arise if the entire industry fails to uphold those standards. "We believe the industry, working with governments, will put the right protections in place," he said. The remarks represent Amazon's response to the debate over slowing AI development that Anthropic has sparked, and signal a different stance from Anthropic's proposal to dial back the pace of development.

Amazon, the world's largest cloud company, has developed its own AI model line called Nova, though the Nova models have yet to reach the competitive tier occupied by advanced models from Anthropic and OpenAI.

Amazon Web Services, Amazon's cloud division, is both a partner and investor in AI companies including Anthropic and OpenAI.


yckim6452@heraldcorp.com