Proliferation of data and improved algorithms are two key contributors to the rapid growth of artificial intelligence (AI) in recent years. From smart devices in healthcare to smart factories, the benefits of AI are already evident across various verticals — and it’s only the beginning.

“We are just scratching the surface of AI and machine learning,” says Michael Zeller, Temasek’s head of AI Strategy and Solutions. “The application of AI is profound, comparable even to the rise of the Internet.” 

He leads AI@Temasek, which offers deep expertise in the field and seeks to make strategic investments to accelerate AI’s deployment and the creation of scalable AI products and solutions within Temasek and its portfolio companies. 

In short, the team wants to apply AI technology to generate better business outcomes and ultimately, shape a better world.

Steps to Becoming AI-ready

Businesses are increasingly aware of the potential of AI. According to an Accenture report, an incredible 84 percent of C-suite executives agree that they need to leverage AI to achieve their growth objectives.

“Most organisations are eager to capture the potential of AI, but stall at the overwhelming set of potential applications and use cases that are realistic and feasible,” Zeller observes.

So what does it take for organisations to really become “AI-ready”?

Before deploying AI solutions, it is key for companies to first identify use-cases where AI can make a meaningful business impact to internal operations or customer experience, says Zeller. “This is often very specific to your industry and depends on your strategy and business model. Focus on use-cases that drive competitive advantages specific to your firm, and don’t compete with the AI R&D scope of Big Tech players,” he adds.

For now, Zeller stresses that AI works best with large amounts of data that can be cleaned and refined to power AI algorithms and applications: “If we subscribe to the notion that data is the new oil, then AI is the new oil refinery.”

As a result, companies need to tie their use-cases to key data assets that they can use to train and deploy AI models. Adequate data infrastructure needs to be established to collect, store and share data in a trusted and secure way. Since data is often spread across departments, data silos have to be broken down and collaboration across data owners and users — the business, data science and technology teams — must be fostered.

Responsible deployment of AI solutions requires a human-centric approach. Companies should ask deliberate questions about how far and how fast AI should be pushed. “We should not ‘de-skill’ humans faster than we can ‘re-skill’ them,” Zeller notes. At the heart of it, businesses need to recognise the value of human experience. That means focusing on broad data fluency and training across the organisation.

“It is critical that it’s not just the data scientists or engineers who understand AI; you also need people at multiple levels of the company to appreciate where AI and data can make a meaningful difference, as well as where it cannot,” Zeller stresses.

Challenges in Deploying AI

Among a company’s key decision-makers, a lack of thorough understanding of AI and its capabilities can also become a critical bottleneck. AI needs to be demystified. Key decision makers must understand what AI can or cannot deliver and how it ties into the broader corporate strategy, Zeller says. Only then can a business identify potential AI solutions and the allocation of required resources, as well as enable smoother collaboration across the organisation.

An area of legitimate concern for companies is the data privacy considerations around the collection of and access to data, across jurisdictions and even across departments. Some industries, such as banking, also struggle with access to tools and open-source libraries, due to data security restrictions.

The rise of AI will raise ethical concerns, such as possible algorithmic or data biases and just how far to go with the use of the data. Companies must address such ethical considerations thoughtfully, Zeller says, as they bring potential reputational risks or broad-based consumer pushback if the AI solutions are deployed irresponsibly.

“We aspire to play the role of an enabler of data collaboration across our ecosystem and tackle some of these challenges,” shares Zeller. 

To do so, Temasek is focusing on AI ethics and governance, which Zeller says “form the foundation for innovative and accurate algorithms, personalised consumer experiences and progress on research”.

Temasek has been an active contributor to Singapore’s Model AI Governance Framework. Deputy CEO, Chia Song Hwee, is a member of the Singapore Government’s Advisory Council on the Ethical Use of AI and Data. At the start of 2021, Temasek also signed up to the Singapore Computer Society’s AI Ethics & Governance corporate pledge to create awareness, propagate the ethical application of AI and encourage staff to adopt good principles and guidelines.

“If we promote the use of AI for the betterment of society rather than feed into the perceived threat, we can help inform, curb fears early, manage expectations better and determine the right ethical approaches to key questions,” says Zeller.

“Data science and AI can do a lot of good — that needs to be our primary focus and guiding principle.”