IBM: Gen AI off to a good start in Singapore

Singapore businesses are warming up to AI and generative AI (gen AI), according to long-time AI investor IBM, which also sees 2024 as the year of adoption for gen AI.

In a panel discussing the implications of AI, Colin Tan, GM & Technology Leader, IBM Singapore, said that he has been seeing strong traction and interest from clients to leverage gen AI to drive productivity and greater efficiency, with some pilots already under way. 

"In the last six months, (the main topic for) every business meeting has been gen AI," he said. "The organisations in Singapore are definitely warming up and waking up to the potential and possibilities of AI."

Tan further observed that generative AI adoption should be based on two core principles:

- An open platform: Is the platform based on the best AI and cloud technologies available, and allow them access to the innovation of the open community and multiple models? 

According to Tan, organisations should not embrace a specific language model, but embrace innovations from the open community. Considerations include the language model, whether there is support for multiple large language models (LLMs), and the different data types that can be used, such as images, text, or code.

 - Trust, the key roadblock and also the key enabler. Tan explained that while outcomes with gen AI are known, what goes on to arrive at the outcome is still unknown.

"We need to have clarity in monitoring models, fairness, openness, traceability and transparency of the data going in and out," he said.

IBM launched watsonx, its data and gen AI platform, in 2023. The platform has been designed with trust in mind. The company has also introduced a collection of generative AI models to advance the infusion of generative AI into business applications and workflows. In September, IBM announced the general availability of the first models in the watsonx Granite model series.

On the services side, IBM Consulting is accelerating generative AI adoption through its expertise and ecosystem partnerships, shared Ng Lai Yee, Singapore Managing Partner & Country Leader, IBM Consulting. She said that IBM Consulting is already helping clients worldwide apply enterprise-scale AI with a human-centred, principled approach, and is approaching telcos and banks in Singapore with the same offer.

Three pillars govern IBM Consulting's offerings, Ng said:

- Customer service is the No. 1 priority for generative AI outcomes. For instance, Bouygues Telecom was able to enhance customer interactions and reduced operational costs by a projected US$5 M a year with IBM AI. IBM Consulting has seen some clients experience up to 70% acceleration in time-to-value with foundation models, as compared to traditional AI approaches.

- Talent productivity will be key as today’s employees are frequently overloaded, juggling a multitude of apps while other tasks remain manual. IBM is 'client zero', and has embedded gen AI in its own HR solutions.

"(We're) now breaking out to help HR transform journeys," Ng said, sharing that there has been a 40% improvement for HR productivity. 

Rather than adding modern applications to outdated systems, organisations can use gen AI to enhance resource utilisation by automating and streamlining parts of the application modernisation process. Sixty percent of IBM’s Ansible Playbook content was automatically generated by watsonx Code Assistant for Red Hat Ansible Lightspeed in a tech preview, for instance.

- Governance and safety is paramount. Organisations that employ AI to unlock new value and insights, accelerate discovery or to gain competitive edge, have a fundamental responsibility to foster trust in the technology.

IBM Consulting announced a global Center of Excellence for gen AI in May 2023. Over 1,000 gen AI experts at the centre are ready to offer deep industry and domain expertise to actively build and deploy watsonx for clients. 

Even during the use of traditional AI, Singapore has been at the forefront of defining AI governance frameworks, but things are moving to the next level.

Puruthoshama Shenoy, CTO, IBM Singapore, said that Singapore has been at the forefront of AI governance, and is now implementing how the principles of AI governance can be put into practice. "It’s a big shift moving from principles to practice," he said.

Shenoy shared that there are two Singapore organisations that IBM is working with closely to drive trusted AI:

- IMDA has set up the AI Verify Foundation to harness the collective power and contributions of the global open-source community to develop AI Verify testing tool for the responsible use of AI. IBM is one of seven premier members who will set strategic directions and development roadmap of AI Verify.

- The Monetary Authority of Singapore (MAS) is collaborating with IBM to demonstrate the MAS Veritas toolkit integration with an IBM AI governance solution that helps to comply with MAS fairness requirements, and at the same time, provide a complete AI lifecycle governance tool for continuous monitoring in operations.

IBM’s promise to build trusted and responsible AI includes:

- The AI Alliance, an international community of researchers, developers and organisational leaders committed to support and enhance open innovation across the AI technology landscape to accelerate progress, improve safety, security and trust in AI, and maximise benefits to people and society everywhere.

Members of the AI Alliance believe that open innovation is essential to develop and achieve safe and responsible AI that benefit society rather than benefit a select few big players. A*STAR, an AI Alliance member, is a key partner for IBM in Singapore.

- Watsonx.governance. The watsonx solution, generally available in November, helps businesses automate responsible, transparent machine learning and generative AI workflows on one, integrated platform. 

Shenoy said watsonx.governance had been designed to address risks associated with AI, such as reputational risk, regulatory risk, and operational risk. "We can help our clients embrace those," he said.

AI governance is a differentiator for IBM. Shenoy pointed out that IBM is the first in the industry to publish information about its training methodology for the foundational model. The company is further providing an IP indemnity (contractual protection) for its foundation models. "We know exactly where the data came from," he noted. 

"We're very confident that.. we are able to indemnify (any output) because we know the exact training process we went through."

If clients want to obtain the best outcomes, a combination of traditional and generative AI is best as each has unique advantages, IBM said. "Governance is key and a differentiator for us," said Agnes Heftberger, GM and Technology Leader for IBM ASEANZK.

"We've been doing that for traditional AI and bringing it to generative AI as well."

The company is bringing some LLM metrics to clients so that they can decide whether a particular model should be deployed. The information will act like a nutrition label, so key decision makers can decide to whether to put specific models in production, and what risks are associated with it.

Open innovation is needed to accelerate gen AI in Singapore, Ng added. 

She observed that innovation has traditionally been closed, or internally focused, used to maintain secrecy, surprise, and competitive advantage. But traditional “closed” innovation is no longer adequate. "Companies...will have 60% higher increase in their revenue if they invest in open innovation," she shared.

IBM has been working with AWS on the general availability of Amazon Relational Database Service (Amazon RDS) for Db2, a fully managed cloud offering designed to make it easier for database customers to manage data for AI workloads across hybrid cloud environments.

• Innovation now is a team sport. It is no longer at the internal company organisational level. Companies need to move from 'plus AI', existing infrastructure with AI added, to 'AI plus', solutions designed with AI in mind.

- In Singapore, many clients are still talking about plus AI, ie. With the current processes, how do you add AI to make it more efficient. There is a need to think AI plus to completely change the way that services are delivered. Companies must consider how to apply open innovation, AI plus, and working across the ecosystem to improve business outcomes.

Looking ahead in 2024: The future of deep tech is dependent on deep trust.

• Deep tech requires deep trust. Organisations need to build trust into everything – from data to people, and from machines to operating models.

• 2024 is the year where we will begin to see organisations move from 'plus AI' to 'AI plus'The wider ecosystem will no longer be a part of the strategy but the strategy. Building a trusted ecosystem will become increasingly central to the strategy.

• An integrated hybrid cloud strategy and robust security strategy are critical.