By Kami Navarro
Though the term “artificial intelligence” (AI) was first coined nearly 65 years ago, it’s only recently that the technology has come of age. Loosely defined as the capability of a machine to do tasks associated with human intelligence, AI is rapidly gaining traction for the enormous potential it holds across diverse industries worldwide.
With its ability to automate standard business processes and generate data-driven insights on a massive scale, 58% of respondents in McKinsey’s latest Global AI Survey report having embedded at least one AI capability into their business unit. Still, some foundational challenges remain in the path of widespread AI adoption.
Case in point: a recent webinar poll from EY indicates that distrust in AI is prevalent in Asia-Pacific, slowing down adoption rates in the region. Three challenges in particular were raised—namely, a lack of transparency, issues with bias and the perceived difficulty in explaining AI concepts to the public.
Singapore-headquartered company BasisAI is determined to tackle each of these challenges head-on by enabling businesses to adopt AI in a responsible and accountable way. For its forward-thinking mission, BasisAI has generated significant buzz—raising over US$6 million in seed funding from Temasek and Sequoia India despite having just come out of stealth mode last January 2019.
Unpacking the black box
AI’s most infamous problem is known as the “black box,” which encapsulates all the challenges previously raised by the EY webinar participants. Despite their increasing sophistication, AI systems are still notoriously opaque—similar to looking into a black box. “There is often very little visibility into how they work and why decisions are made,” explained BasisAI co-founder and CEO Mr Liu Feng-Yuan.
At its core, BasisAI strives to demystify this black box of AI. “Systems should be trusted and explainable,” said Mr Liu. “We wanted to help supercharge machine learning capabilities in enterprises more quickly—but in a responsible way, with AI engines that are trusted by end users.” This shared vision led Mr Liu to take a leap of faith and start BasisAI with twin brothers Linus and Silvanus Lee.
With over 20 years of experience in leading data and engineering teams spread across the three co-founders, the company’s leadership team packs a powerful punch. Mr Liu, for instance, was the former chief data scientist of the Singapore government. Not to be outdone, Mr Linus Lee was one of Twitter’s first data scientists while Mr Silvanus Lee was Uber’s Director & Head of Product Data Science.
Building the bedrock of AI
BasisAI’s flagship product is a cloud-based platform-as-a-service called Bedrock. As suggested by its name, Bedrock is envisioned by the co-founders to be the foundation on which enterprises can responsibly deploy AI in a real-world setting.
In Mr Liu’s experience, data science and machine learning teams in digital organisations typically work in silos, compromising their ability to deliver robust AI systems. Bedrock therefore brings the two teams together by giving them a secure means to collaborate while retaining control over their own data and code. It also automates machine learning workflows, accelerating the productivity of the engineers and organisation as a whole.
With its built-in AI governance capabilities, Bedrock also seeks to turn the infamous black box into a translucent glass pane. It does so by providing visibility on all parts of the workflow, creating a digital trail that could audit predictions made by AI software. In this manner, organisations can ensure that their models are fair and easily explainable—building their user’s trust in the model along the way.
Though external solution providers can provide generic, pre-trained AI models, an internal lack of knowledge in the organisation can also create a black box. This is why BasisAI helps business leaders build internal capabilities through exploratory use case sessions and even system design workshops. “It’s not about throwing AI at the problem,” explained Mr Liu. “[It’s about] helping customers improve their experience and optimise the way they run their operations in a more data-driven way.”
Towards an AI-enabled future
One of BasisAI’s biggest selling points is their ability to conceive and execute a live AI engine in a matter of months—and even weeks. According to Mr Liu, this is made possible by the close cooperation of the team with their customers from the conceptualisation phase all the way to deployment. Bedrock is also easily integrated into existing systems, he added.
It may be early days yet, but Bedrock has already been successfully deployed in several instances. It has helped a low-cost carrier adopt a dynamic pricing model by processing a deluge of data, ranging from historical booking and route information to schedule changes. It’s also been used to build hyper-relevant personalised search models to increase a marketplace startup’s customer engagement.
Having recently accredited by IMDA, BasisAI can now be reliably harnessed by government agencies and enterprises. AI, in particular, could revolutionise public services by making them more personalised and anticipatory. Machine learning could also assist in the processing of high-volume regulatory cases and even extract relevant information for national security surveillance.
Currently, the BasisAI team is working to add features that will enhance user trust in the AI systems built by Bedrock. “Our mission has always been about the responsible use of data and development of trusted systems,” said Mr Liu. As more and more businesses begin to leverage AI and machine learning, BasisAI will certainly play a critical role in ensuring the responsibility and accountability of AI as a whole.
This feature is the first in a series of articles profiling accredited companies under IMDA’s Accreditation@SG Digital (Accreditation@SGD) programme. First launched in July 2014, Accreditation@SGD contributes to an innovative infocomm media ecosystem by accrediting promising Singapore-based tech product companies to establish their credentials, build business traction, and help them to grow and compete in the global market. The evaluation process provides an independent third party evaluation on the SGD-accredited companies’ claimed product core functionalities and ability to deliver.
is the second in a series of articles profiling accredited companies under IMDA’s Accreditation@SG Digital (Accreditation@SGD) programme. First launched in July 2014, Accreditation@SGD contributes to an innovative infocomm media ecosystem by accrediting promising Singapore-based tech product companies to establish their credentials, build business traction, and help them to grow and compete in the global market. The evaluation process provides an independent third party evaluation on the SGD-accredited companies’ claimed product core functionalities and ability to deliver.
As of February 2020, over S$430 million worth of project opportunities have been created for accredited companies. Close to 1,000 projects have also been awarded. For more information, please refer to this link: www.imda.gov.sg/accreditation
.webp)