The Edge of (AI) Glory: Why Companies Need Cyber Resilience Strategies When Adopting Edge AI
Andy Ng, Vice President and Managing Director for Asia South and Pacific Region, Veritas Technologies
While this may have been unimaginable once upon a time, the use of edge computing and artificial intelligence (AI) is now ubiquitous among the various consumer goods industries that improve our day-to-day lives. Edge AI is increasingly moving intelligent computation from cloud-based models to locations closer to data sources. This shift is driven by the benefits of reducing network traffic, enabling near-instant decision-making in time-sensitive applications like manufacturing and autonomous systems, and enhancing privacy through local data processing.
Today, the likes of consumer electronics, healthcare, retail, and smart homes, combine the advantages of real-time AI-powered data analytics with the hyper-connectivity of the Internet-of-Things (IoT) to draw insights from numerous devices. For instance, health monitors can help find anatomical anomalies through wearable sensors, that are otherwise too difficult to be observed by the naked eye.
Apart from consumer products, the technology has also become increasingly prevalent in industrial settings that contribute to more cost-efficient production. Accelerated manufacturing lines enable quick and seamless identification of parts and components, highlighting the increasing importance of edge AI. Additionally, micro, thin and thick-edge AI can also facilitate model training or retraining for scenarios that might involve sensitive data.
In Singapore, the adoption of Edge AI aligns with the city-state’s Smart Nation initiative, which aims to harness technology to improve urban living, create economic opportunities, and build stronger communities. For instance, Edge AI can enhance public safety through advanced surveillance systems capable of real-time anomaly detection, and it can optimise public transport systems by analysing traffic patterns to reduce congestion. With estimates that Singapore’s AI sector could see its market volume reach $US4.64 billion by 2030, business leaders must take advantage of edge AI’s capabilities, generating meaningful insights that simultaneously safeguard and strengthen company data.
Challenges brought by edge AI – complexity created from storage in edge AI devices
Like any form of innovation, businesses must remain vigilant about the numerous challenges posed by the implementation of edge AI technology. With the growing amount of data that people generate daily, data storage for devices leveraging edge AI technology has become increasingly challenging, as models require numerous data points to produce accurate insights.
For AI on the edge to work, there are new requirements for decentralised intelligence – with potential challenges centred around power, data throughput latency, accuracy, safety, security, regulations, diverse workloads. Running complex AI models on edge devices can be also challenging due to the high processing power required. While a body of data that pulls different applications into its orbit can create better analytics, the cost of hosting big data can prove financially and technically challenging for any business. Reaping the benefits of data gravity can quickly manifest into increased administration costs and inadequate security measures that leave organisations at the risk of exposure.
Hosting multiple work streams will require proper monitoring to ensure optimal performance and avoid unnecessary manual work. But the computational complexity of deploying numerous edge AI models across diverse locations can make it hard to protect and oversee a large scale of running data.
Reinforcing cyber resilience measures as prevention and mitigation
The critical need for edge AI to be paired with strong security and cyber resiliency measures cannot be understated. Creating routine cyber checks will ensure that data is safely stored, mitigating any operational disruptions that might occur. Companies should constantly monitor and detect data breaches happening with their edge AI systems and have a recovery plan in place.
A best-practice strategy might prove pivotal in isolating any potential breach, ensuring that other day- to-day activities can resume operations seamlessly. A cyber resilience strategy that identifies both the type and sensitivity of the data that edge AI programmes monitor will also help businesses apply edge AI more effectively. Frameworks such as a Zero Trust policy are also effective in ensuring that devices are not being manipulated by malicious actors.
The adoption of edge AI technology can come with many benefits for any business. Nonetheless, its full capabilities can only be unleashed with proper safeguards that protect routine operations. In Singapore, where technology is a cornerstone of economic and social inclusivity, building cyber resilience is crucial to realising the full potential of Edge AI. By adopting cyber best-practices and understanding the challenges that may manifest from unsupervised edge AI activity, companies are well placed on the edge of (AI) glory.
