Navigating Tomorrow: AI's Role in Shaping the Future of Work and Industries

In the realm of technological evolution, Artificial Intelligence (AI) stands as a multifaceted force set to redefine industries and reshape the very landscape of work. AI is often reduced to a mere buzzword, says Lee Hickin, AI Policy and Technology Advocacy Lead at Microsoft Asia, who was also a keynote speaker at the Malaysia Digital Expo 2023 (MDX 2023) held in Kuala Lumpur.

“It's essential to recognise that AI encompasses various technologies, specifically computational, enabling the performance of tasks with a degree of intelligence and predefined objectives,” Hickin says. To illustrate, AI can excel in identifying minuscule anomalies within thousands of material compounds crucial for constructing infrastructure, ensuring heightened safety and durability. 

Additionally, it can play a role in aiding administrative assistants in large corporations by summarising a month's worth of purchase orders and providing insights into the primary areas of business expenditure. Despite their disparate applications, both instances underscore AI's extensive and diverse nature.

"It is a tool that will transform the way we work, allowing us to consume, understand, and learn from significantly larger amounts of information," he states.

Interaction with data & knowledge
Hickin highlights a fundamental value that AI brings to humanity – a new interface into data and knowledge. Unlike the historical barriers requiring individuals to comprehend complex computer languages, generative AI and large language models have birthed a natural language interface. Users can now extract knowledge from data simply by asking questions in everyday language, eliminating the need for specialised technical skills.

The practical applications of AI are virtually limitless, spanning healthcare, infrastructure, education, science, government, and beyond. Hickin emphasises the need for a nuanced consideration of safety and risk during AI's application. While the technology promises immense value, the deployment phase, where AI interfaces with human lives, necessitates deep investment in control, safety mechanisms, and risk mitigation regulations.

“I would stress that AI development and research should be done with safety in mind. It is the point at which we deploy AI such that it touches human lives that we need to invest deeply in control, safety brakes and risk mitigation regulation.”

Controlling AI functions

Hickin articulates a three-layered approach to controlling AI functions, beginning with the developmental phase. Here, the emphasis lies on safe training data, legally acquired information, and the security of model data. The next layer involves deploying AI into applications or services, focusing on minimising risks and establishing trust between deployers and users. The final layer resides with end-users, where intelligent UI/UX designs and application services guide users in effectively and safely interacting with AI.

Addressing the reliability of AI outputs, Hickin asserts its dependence on user input. If users pose well-formulated questions, they are more likely to receive reliable answers. Embedded and retrieval augmentation tools can assist the AI model in providing a more grounded response to the user. It is often more reliable and something the user can validate outside of the AI because it will give a citation or source for its output.

Malaysia's AI opportunity
As to Malaysia's preparedness for the AI opportunity, Hickin notes that Malaysia, with its highly digitally aware workforce, a substantial working-age population, and a keen interest in integrating AI into business practices, holds the potential to embrace AI responsibly.

“It would be inappropriate and unfair for me to comment on a nation’s overall preparedness. I would say that Malaysia is no less prepared than any other country.

“Malaysia has in front of it the opportunity to take, embrace and enhance the lives of its citizens through the responsible adoption of AI,” he underscores.