The headquarters of the Nationwide Intelligence Service in southern Seoul / Yonhap
The Nationwide Intelligence Service (NIS) stated Thursday it has issued a joint steering on synthetic intelligence (AI) provide chains along with six cybersecurity companies from like-minded nations, together with the USA, Japan and Australia.
Within the joint steering, titled “Synthetic intelligence and machine studying: Provide chain dangers and mitigations,” the NIS and companion companies emphasize that AI techniques ought to combine safety issues from the design stage reasonably than treating safety as one thing to be addressed solely after deployment.
It additionally outlines potential dangers throughout six key elements of the AI provide chain — AI coaching information, mannequin, software program, infrastructure, {hardware} and third-party providers — whereas presenting really helpful mitigation methods for every space.
In accordance with the NIS, the steering was ready amid rising considerations that the rising complexity of AI provide chains may introduce safety vulnerabilities, together with the opportunity of hidden backdoors or malicious code embedded in techniques.
The NIS stated it’ll proceed cooperation with worldwide companions to help the secure utilization of AI applied sciences.
