The US, Britain, and over a dozen other nations unveiled the first detailed international agreement on how to keep AI safe, pushing for firms to create AI systems that are “secure by design”.
The 20-page document unveiled on Sunday said that the 18 participating countries agreed that companies designing and implementing AI were required to develop and deploy it in a manner that protected customers and the wider public from misuse.
The agreement signed by the participating nations is non-binding and features mostly general recommendations like monitoring AI systems for abuse, vetting software suppliers, and protecting data from tampering.
Nonetheless, Jen Easterly, the US Cybersecurity and Infrastructure Security Agency’s director said that it was imperative that so many countries put their names to the idea that AI systems desired to place a primary focus on safety.
The agreement is the most recent of several global governments' attempts, most of which lack teeth, to influence artificial intelligence (AI) development, which is becoming more and more significant to business and society at large.
Germany, Italy, the Czech Republic, Estonia, Poland, Australia, Chile, Israel, Nigeria, and Singapore are among the eighteen nations that have ratified the new guidelines, alongside the United States and Britain.
The framework addresses concerns about how to prevent hackers from taking advantage of AI technology and offers suggestions like releasing models only after carrying out sufficient security tests.
