Understanding of AI’s Benefits and Consideration of AI Ethics and Governance are Keys to Broadening AI and High Performance Computing Adoption in Singapore

 

Artificial Intelligence, or AI, is commonly understood as the study and engineering of computations that make it possible to perceive, reason, act, learn and adapt.

AI is consistently cited as a key technology area with the potential to affect all aspects of the digital world. It is also one of the four frontier technology focus areas identified by IMDA to lay a strong foundation for information communication media for Singapore1.

Gartner’s 2020 emerging technology hype cycle identified 30 technologies that show promise in delivering a high degree of competitive advantage over the next five to 10 years. AI-related technologies feature strongly in this line-up2.

In Singapore, the AI market has the potential to become a US$960 million market in 2022 and US$16 billion by 2030. AI here includes a wide range of technologies used to “analyse, organise, access and provide advisory services based on a range of unstructured information”3.

 

Misconceptions and insufficient knowledge are hurdles to broader adoption of AI

Financial values aside, the deeper value in deploying AI lies in its potential to improve our quality of life. However, due to insufficient knowledge and misconceptions about AI, the adoption and application of AI in Singapore are not as widespread as it could be.

Business leaders hesitate to adopt AI in their business because they are not familiar with the use cases and how those can contribute to the company’s bottom line. Many leaders are mindful that employee training to engender new skills is needed to make the technology adoption effective. As a result, they also hesitate to consider AI because they are not informed about the skills required of their employees, and may tend to assume the training needs to be complex. Without a view of the total cost of deployment of the new technology, the consideration for adopting AI stops there.

Furthermore, many employees hold the following misconceptions, which causes them to have misgivings about the deployment of AI in their workplace:

  • AI will take over their jobs. 
  • Employees must become programmers or data scientists to continue to be employed when AI is deployed.
  • The skills needed to be AI-relevant are a narrow set of programming and data science skills, which they may not feel capable of mustering. 

AI takes over tasks, not jobs. Hence, it is true that jobs that consist of single tasks are at risk. In reality, however, most job roles can be enriched when AI is deployed to take over mundane tasks. Also, new jobs in roles supporting AI deployment and maintenance are created.

There is an opportunity to mobilise our workforce to become AI-relevant by turning around these misconceptions and broadening the understanding of the reality of AI deployment.

In these challenging times, we can inspire hope in our collective future when individuals can see that through AI, existing roles can be improved and there are real, possible paths to opportunities for new job roles that go beyond programmers and data scientists.

In tandem, when we can show business leaders how their existing workforce can pick up AI-relevant skills to use AI tools and systems in their jobs, the leaders can better visualise the business case for AI. They may become more willing to consider use cases in their business.


Roles for stakeholders bridge the gap

To correct the misconceptions so that all parties have the right understanding of AI and its possibilities, SGTech advocates the following actions by stakeholders.

Firstly, individuals need to gain an understanding that AI-related jobs are not limited to data scientists and programmers and that they can (a) easily “plus-skill” 4 themselves with AI-relevant skills to leverage AI tools and systems that can help them in their work and (b) consider opportunities in roles supporting the deployment of AI. To achieve this requires a broad-based public outreach, possibly taken by IMDA.

Secondly, to supplement the broad-based understanding of possibilities from using AI tools, technology and systems at work, the unions can play a part to facilitate workers to learn about sector-specific tools and encourage them to “plus-skill” themselves to take advantage of opportunities.

The third segment is the trade associations and chambers (TACs). Focussing on business owners and leaders, TACs can encourage and facilitate their members to explore how the use of AI can increase the productivity and motivation of their employees and make an impact on their bottom line.


SGTech initiatives to showcase the benefits of AI through the lens of business leaders and adopters/workers

SGTech strives to drive the adoption of AI among Singapore companies to improve our quality of life. We want to do this in a phased approach, focusing on specific industry sectors at a time.

  1. Work with like-minded champions, such as AI Singapore (AISG) and the unions (through NTUC) of identified sectors to reach out to the workforce. This outreach is to show how AI tools, technology and systems can help in their work. For example, administrative staff can be relieved of mundane clicking for monthly submissions when the company uses robotic process automation such as the open-source TAGUI tool by AISG.
  2. Work with TACs of the identified sectors to reach out to business leaders in that industry, to demonstrate the business case by showing how AI can be used to achieve better results with the same set of employees. For example, a cleaning company was able to win and fulfil more contracts with their current employees by deploying cleaning robots and having their employees undergo training to operate the robots. After the training, the employees received a salary increment, and they were also relieved of the mundane cleaning duties. 
  3. Curate and recommend relevant and quality training programmes (both online and instructor-led) through which individuals can “plus-skill” themselves.
  4. Build a sustainable talent pipeline for AI-related opportunities that will continue to grow. Work with SGTech members and their customers/users as well as Institutes of Higher Learning to expand a broader definition of AI-related jobs (including those supporting and maintaining AI deployments), to encourage the exploration of such opportunities. 

At the same time, we want to facilitate continuous development in AI in the tech space to sustain a pipeline of cutting-edge and useful AI-based solutions in Singapore. 

SGTech will keep our members, the AI solution builders and developers, apprised of the latest technology developments, to leverage the best technology to create solutions that meet society’s needs and improve our quality of life. We will build platforms to facilitate technology sharing and discussion.

Through the above initiatives, SGTech wants to create the conditions to facilitate broader adoption of AI applications among companies in Singapore.

 


Greater awareness of data ethics and governance for AI

Trust is a crucial factor underpinning broader consumer adoption of AI technology.

Data is the foundation of AI. Organisations that build AI tools and systems must take the lead to understand the issues related to data ethics and governance and adopt practices that give these issues due consideration. This will engender trust in AI tools and solutions.

Companies that build and use AI tools and systems should be aware of the considerations and good practices for ethics and governance of data collected and used for AI. 

The PDPC and IMDA have published the Model AI Governance Framework5, providing detailed and readily implementable guidance to address key ethical and governance issues when deploying AI solutions. However, neither the framework nor the issues are top of mind for many companies using or considering AI.

To bridge this gap, the system integrators and solution providers helping the user companies, as the subject matter experts in AI, should ensure that they are familiar with the issues so that they can take the lead and remind their users to encourage compliance.

At the other end of the spectrum, we urge government agencies promoting the adoption of AI, such as IMDA and MAS (financial solutions), to raise awareness among users of AI of the need for proper consideration of the implications of the AI data and model used.

 


SGTech initiatives to promote awareness of AI data ethics and governance among tech companies, especially builders of AI solutions

As the association representing the tech industry, SGTech aims to raise awareness among builders and users of possible risks resulting from insufficient consideration of the implications of data collected and used for AI. We think that this is best achieved by promoting thorough debate and discussion to facilitate the exchange of views and sharing of possible perspectives and experiences.

To this end, SGTech initiatives would include:

  1. Organising workshops to facilitate AI solution builders to explore and consider the possible (unintended or maliciously intended) implications when the AI model is not set up correctly, e.g. bias in models leading to unintentional discrimination in the provisioning of services6. The intended outcome is for builders to be aware of possible risks and consider including sufficient checks and balances corresponding to the extent of the risk.
  2. Providing or directing members to resources that can guide them to consider the issues relating to data so that they can define their policies and practices accordingly.
  3. Facilitating discussions (e.g. with SGTech’s Data Protection Committee) among members on challenges faced, such as in implementing according to guidelines, then surfacing them to relevant agencies, e.g. PDPC for deliberation as necessary.

When the industry and ecosystem consistently consider possible risks in AI models and adopt good practices to include checks and balances corresponding to the risks, we can collectively engender greater trust in AI solutions among users (both individuals and companies). The increased confidence will pave the way for greater adoption of AI.


Conclusion

For Singapore to derive benefits from AI, we need to create the conditions to facilitate broader adoption of AI tools, systems and technologies among businesses, by demonstrating the business case and building the talent pipeline with AI-relevant skills beyond data science and programming. Adoption can be further enhanced when AI solution builders develop their awareness of potential underlying ethics and governance issues and, as subject matter experts, adopt good practices to mitigate the resultant risks for the users. SGTech sees the importance of playing an active role in this process.

 

About

AI & High Performance Computing is one of the chapters at SGTech that offer strategic support to the industry. The Chapter aims to foster an integrated ecosystem to accelerate the development of AI capabilities among tech companies, adoption of AI among enterprise users and build a sustainable talent pipeline to support these efforts.

Please contact wanxin@sgtech.org.sg if you’d like to find out more about the Chapter.

 

Footnote:

"Singapore's Digital Economy Gunning for Four Key Areas." IMDA, Jun 2017. https://www.imda.gov.sg/news-and-events/impact-news/2017/06/singapores-digital-economy-gunning-for-four-key-areas 

https://www.gartner.com/en/newsroom/press-releases/2020-08-18-gartner-identifies-five-emerging-trends-that-will-drive-technology-innovation-for-the-next-decade

The Future of Services, IMDA. https://www.imda.gov.sg/-/media/Imda/Files/Industry-Development/Infrastructure/Technology/Technology-Roadmap/Annexes-A-4-Artificial-Intelligence-and-Data-and-Blockchain_Full-Report.pdf?la=en 

The term “plus-skill” is used deliberately, instead of re-skill or up-skill. It recognises the importance and significance of employees’ current skills and domain knowledge, while signalling that the AI-relevant skills are an add-on and a way to enrich their jobs.

https://www.pdpc.gov.sg/Help-and-Resources/2020/01/Model-AI-Governance-Framework

Depending on the AI application, inherent biases can have far-reaching implications in actual use.To take an extreme example, an AI visual assessment tool to triage patients arriving at a hospital emergency department could lead to severe consequences if there were a bias that was unknown or unconsidered.

---

Published Nov 2020