How can Malaysia ensure AI systems remain controllable if they malfunction?

Question

Accepted Answer

Malaysia's Strategy for Ensuring Controllability of AI Systems During Malfunctions

Malaysia's approach to ensuring AI systems remain controllable, even when they malfunction, centres on rigorous pre-deployment assessments and the implementation of robust fail-safe mechanisms. Digital Minister Gobind Singh Deo has stressed that thorough safety and security evaluations are paramount before any new technology, including artificial intelligence, is adopted and deployed. This proactive stance aims to prevent operational failures and maintain user trust by ensuring that inherent risks, verification processes, and control mechanisms are fully understood and addressed prior to widespread implementation.
Pre-Deployment Safety and Security Assessments

A foundational element of Malaysia's strategy involves conducting comprehensive safety and security assessments before new AI technologies are introduced. This process is designed to identify potential vulnerabilities and operational risks associated with AI systems, ranging from robotics to Large Language Models. The goal is to ensure that any technology adopted is secure for widespread use, thereby safeguarding user reliance and preventing unexpected failures.
Integration of Fail-Safe Layers and Robust Infrastructure

The emphasis is placed on the necessity of incorporating robust digital infrastructure and multiple fail-safe layers within robotics and autonomous AI deployments. These layers are critical for maintaining control even if an AI system encounters an issue. The principle is to ensure that if any part of the system goes wrong, there are built-in redundancies and backup systems that allow the situation to remain manageable and under human oversight.
Understanding Risks and Verification Processes

Users and policymakers are urged to gain a thorough understanding of the inherent risks associated with AI technologies, alongside the necessary verification processes and control mechanisms. This educational aspect is crucial for informed decision-making regarding AI adoption. By comprehending these factors, stakeholders can better anticipate potential problems and support the implementation of appropriate safeguards.
Learning from Past Incidents

Lessons learned from past high-profile AI event launches, which have experienced technical glitches such as issues with autonomous robots, underscore the importance of technological readiness and security controls. These incidents serve as critical reminders of the need for meticulous planning and the implementation of effective control measures to prevent similar disruptions and maintain public confidence in AI technologies.

Reading history

Hide →

This history is cleared when the page is closed.

How can Malaysia ensure AI systems remain controllable if they malfunction?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends