Is Singapore AI ready?

Artificial intelligence (AI) anticipation is high. At the Innovfest Unbound Conference, discussion was brimming with the possibilities that rising digitalisation will create; with smart technology expected to transform customer experience and power vast business growth.

But what about current implementation?

Across Asia Pacific only 41% of companies are using AI, and figures in Singapore are even lower: less than one third (24%) of sales teams have put AI systems into action. It seems limited understanding of how to leverage intelligent tools is leading firms to hold tight; 37% are waiting to see what benefits AI will bring before investing.

As the fourth industrial revolution progresses, firms must become high-tech to survive. However, it’s important to get data fundamentals right before launching AI capabilities. If data is fragmented, unclean, or contains human bias, businesses put its technology – and business growth – at risk. As recently noted by Victor Lim, Vice President of IDC Asia Pacific: “to succeed in the AI race, markets in the region need to substantially improve their readiness.”

AI adoption: where are we now?
Often ahead of the curve on tech innovation, it's no surprise Singapore's AI efforts stretch back years. The government partnered with Microsoft in 2016 to explore chatbot integration for public services and invested the better part of $2.82b in digital infrastructure as part of its Smart Nation programme. Last year, AI Singapore (AISG) made multiple steps towards increasing usage; including launching two initiatives to increase national capability: AI for Everyone (AI4E) and AI for Industry (AI4I). At this year's Davos, the government released its Model Framework for AI governance; the first ethical AI guidelines in Asia.

For all this development, however, Singapore is not leading with implementation. According to recent IDC research, Singapore ranks third (9.9%) for regional AI adoption; far behind major force Indonesia (24.6%). High-level eagerness to embrace AI isn't translating into everyday use, and the likely cause of this gulf is uncertainty. Though positive, the Model Framework is limited on finer details of how to keep intelligent machines under control, leaving many companies unsure how to proceed safely.

Next steps: powering the AI evolution
Asia Pacific is rapidly becoming a hotbed of AI activity; with spending on AI systems due to hit almost $7.5b (US$5.5b) in 2019 – with new players such as China set to lead the way. Yet, before employing advanced technology companies must clean and streamline their data and ensure fragmented data is connected across the organisation, and complies with any regulatory frameworks before launching new AI initiatives, to ensure ethical application. Furthermore if Singapore is to maintain its market position, companies must pick up the pace of AI execution, but developing a robust preparation strategy is essential and should cover the three key pillars of ethical AI:

1. Play by all the rules
The advantages of following official standards don't stop with staying on the right side of the law, and avoiding fines. By complying with the highest current benchmarks, companies can gain a useful initial idea of how AI systems should be structured to ensure advanced tech doesn't run off course. Plus, voluntarily going beyond minimum requirements shows a public commitment to ethical procedure that fuels consumer confidence; helping firms improve the impact of insight-driven marketing and services. As a result, stage one is implementing local AI and data guidelines – including Singapore's ethical model and new Trusted Data Sharing Framework – and regional initiatives such as the APEC Cross-Border Privacy Rules (CBPR).

2. Keeping watch on machine bias
AI algorithms are only as reliable as the data feeding them. That’s why stage two is building on optional frameworks with additional measures, centered around keeping data accurate and free from factors that can skew output, such as bias. One of the most effective approaches companies can take is applying smart data orchestration. By collecting, amalgamating and cleansing data from multiple sources in real time, orchestration platforms transform data into a unified insight hub that's easier to oversee, and instantly filtered to remove hazardous bias. Of course, even intelligent data coordination can't entirely eliminate all possibility of human partiality, which means extra vigilance is essential; including advocating diversity in the teams programming AI and subjecting resulting analysis to regular, granular scrutiny.

3. Protecting the access cycle
Consent and transparency have long been part of Singapore's data landscape, with the data bill in force since 2012. As global shifts such as Europe's General Data Protection Regulation (GDPR) place a brighter spotlight on data rights, consumers are increasingly aware of the value their data has for companies; both as a catalyst for personalized customer experience and AI development. Stage three involves leveraging data best practice to forge closer consumer bonds. For example, with a consolidated pool of consensually given information, firms can serve consistently relevant messages that match privacy preferences across channels, and increase their chances of retaining both consumer trust and data access.

Singapore is certainly keen to join the AI era and reap the rewards enhanced automation and efficiency will deliver. But for now, it remains stuck in the initiation phase, with businesses waiting to see how disruptors fare before taking the plunge. By preparing and focusing on the vital pillars for AI success, businesses will remain at the forefront of tech innovation. By appreciating and following guidelines with their own data security measures, companies can ignite their journey to an AI-centric future.