As nations better understand the transformative power of generative artificial intelligence (GenAI), Vietnam stands at the brink of a substantial opportunity. With its vibrant economy, youthful population, and enthusiastic uptake of cutting-edge technology, Vietnam is an ideal environment for the integration of GenAI, fostering an optimal landscape for its development and implementation. 

According to the Finastra Financial Services State of the Nation Survey 2023, Vietnam stands out as the frontrunner in terms of enthusiasm for generative AI, with a remarkable 91% of surveyed individuals expressing interest in the technology. This percentage is the highest among all the markets included in the survey. 

Statista Market Insights also projects that the Vietnamese generative AI market is anticipated to reach USD 153.80 million in 2024. The forecast suggests an annual growth rate (CAGR 2024-2030) of 23.20%, leading to a market volume of USD 537.70 million by 2030. 

Despite the notable expansion of generative AI, executives are approaching its adoption with caution. A recent study by IBM IBV reveals that 84% of executives consider cybersecurity risk to be the primary obstacle hindering the widespread adoption of generative AI. As the regulatory landscape and AI governance frameworks are still evolving, organisations are taking proactive measures by seeking reliable technology partners to assist them in developing and implementing secure frameworks for the protection of machine learning models. 

Generative AI poses a security risk as threat actors may exploit it to manipulate predictions or deny services. This potential threat makes it imperative for organisations to enhance cyber defence to protect proprietary language, models, data, and generated content, mitigating risks associated with deploying generative AI. 

Furthermore, generative AI applications can increase data and privacy risks due to the reliance on large language models and the generation of new data. This introduces vulnerabilities like bias, poor data quality, and risks of unauthorised access. Therefore, deploying GenAI requires careful consideration and robust measures to address and mitigate these concerns, ensuring data and privacy protection throughout the AI applications’ lifecycle. 

The apprehension regarding generative AI-generated threats is particularly heightened in the context of Vietnam, given the ongoing cybersecurity challenges faced by the country. The National Cyber Security Centre (NCSC) reported a significant surge in cyberattacks in 2023, recording a notable 13,900 incidents. This alarming statistic indicates a worrisome increase of 9.5% from the previous year, positioning Vietnam as the third highest in Southeast Asia for the number of cyberattacks.  

In the current landscape in Vietnam, how can organisations enhance their protection against generative AI-related threats? What strategies and effective measures should they adopt to safeguard the usage of generative AI and become an AI-ready organisation? 

Securing AI Training Data and Advanced Detection to Secure AI Models 

Implementing robust encryption protocols and access controls will further fortify the defence against unauthorised access, ensuring the integrity and confidentiality of the AI training data. Additionally, establishing clear protocols for incident response and continuous monitoring is crucial to swiftly address any potential security threats to the AI training data. 

Moreover, implementing advanced anomaly detection algorithms is essential to secure the usage of AI models by identifying potential data or prompt leakage. Deploying real-time alerting mechanisms for evasion, poisoning, extraction, or inference attacks also strengthens the overall defence against malicious activities, ensuring the strong protection of AI systems. 

Behavioural Defences and Multi-Factor Authentication 

To fortify defences against emerging threats, behavioural defences and multi-factor authentication can be employed to safeguard against new AI-generated attacks, including personalised phishing, AI-generated malware, and fake identities. By incorporating these proactive security measures, organisations can bolster their resilience and effectively mitigate the evolving landscape of AI-driven threats, ensuring a robust and adaptive security posture. 

OpenGov Breakfast Insights are concise, to-the-point, strategic-level discussions designed to bring learning to the highest level! The unique proposition of an OpenGov Asia Breakfast Insight is the integration of cutting-edge inputs from expert speakers and interactive discussion among the participants.