Eighteen countries, including the United States, United Kingdom, and Australia, have released global guidelines to enhance the cybersecurity of AI models, emphasizing the need for models to be “secure by design.” The guidelines recommend measures such as monitoring for tampering, training staff on cybersecurity risks, and maintaining tight control over the infrastructure. However, contentious issues like image-generating models and data collection methods were not addressed. This initiative follows other government efforts to regulate AI, including the European Union’s AI Act and President Joe Biden’s executive order on AI safety and security.
read more