Amazon has waded into the growing debate over artificial intelligence safety, calling for “rigorous testing” and stronger safeguards before new AI models are released.
It said developers should only launch models once they are safe and fully tested, but stopped short of backing calls for the industry to slow the pace of AI development.
An Amazon spokesperson said: “We don’t see it as a choice between progress and safety,” adding that models should undergo “rigorous testing and strong safeguards” before release.
They said said Amazon believed the technology industry should work alongside governments to develop appropriate protections as increasingly powerful AI systems are rolled out.
Amazon’s comments follow widespread calls from some of the biggest names in AI for a more measured approach to development amid mounting concerns about the risks posed by increasingly autonomous models.
Leaders at OpenAI, xAI, Google DeepMind and Microsoft have backed a more cautious approach after Anthropic chief executive Dario Amodei called for the industry to slow the pace of development.
Amazon, however, has not said it will support similar measures, arguing instead that progress and safety can be pursued alongside one another.
The debate is particularly significant for Amazon as it ramps up investment in AI across both its retail and technology businesses.
The group develops its own AI models through Amazon Web Services and its artificial general intelligence division, with the technology helping to power products including its Alexa voice assistant.
Its AWS business is also a major infrastructure provider to AI companies, while its Bedrock platform gives businesses access to models developed by a range of AI firms.
Amazon had until now largely stayed out of the wider debate over the potential risks of advanced AI, as technology companies face growing pressure to set out how they plan to balance rapid innovation with safety.
Click here to sign up to Retail Gazette‘s free daily email newsletter
