(MENAFN- Khaleej Times) There is potential for AI to augment the current capabilities of existing security technologies and skillsets of security professionals, says Dr. David Hardoon, Chief Data Officer, Monetary Authority of Singapore.
Excerpts from an interview:
The Monetary Authority of Singapore recently announced a $27 million grant to promote Artificial Intelligence and Data Analytics in the financial sector. How did this come about?
Data analytics can help organisations enhance processes, unlock insights, and facilitate better decision-making. Equally important is the human resources in an organisation. MAS launched the S$27 million Artificial Intelligence & Data Analytics (AIDA) Grant in November last year to support financial institutions as they embark on new Artificial Intelligence (AI) and data analytics projects. At the same time, we want to encourage them to take proactive steps to equip their workforce with the skills needed for the use of these new technologies. There are two tracks under the AIDA Grant.
Under the Financial Institution Track, the AIDA Grant supports Singapore-based financial institutions that leverage on AI and data analytics techniques to generate insights, formulate strategy, and assist in their decision-making. These techniques may include machine learning, natural language processing or text analytics, deep learning or neural networks, predictive and prescriptive analytics. A key criterion for obtaining the AIDA Grant is that financial institutions will need to consider the impact of the AI or data analytics project on their workforce and develop appropriate training programmes. This could include up-skilling staff with new data analytics capabilities, or re-skilling staff who may be redeployed into new roles.
The AIDA Grant also aims to foster applied research in AI and data analytics. Under the Research Track, the AIDA Grant will co-fund both Singapore-based and global research institutions' AI or data analytics research projects, which have clear applications for Singapore's financial sector. In addition, periodic calls for proposals on specific AI or data analytics topics that benefit the industry will be conducted. There are currently a number of open calls for proposals on the MAS' website (http://www.mas.gov.sg/News-andPublications/Requests-For-Proposal.aspx).
While the power of data analytics is beyond doubt, why do you think a lot of organisations are not able to make the most of it? Where do you think they lack?
The basic building block of data analytics - data - may very well be the stumbling block for many organisations. Organisations often find themselves managing a variety of data repositories. Data integration would then be necessary to bring together the data residing in the different repositories. The issue gets more complex when the different data repositories in use are unable to 'talk' with each other. Absence of or poor processes in place to manage data storage leading to poor data quality is also quite prevalent in many organisations. In an extreme example, an organisation's data may be stored in spreadsheets in individual staff accounts and end up being recorded differently. A substantial amount of effort would be required to clean up and bring these data sets together, before they can be used.
Managing change and aligning expectations and priorities are also important. Data is an enterprise asset that cuts across organisational units of a company. Misaligned expectations and priorities will hinder data initiatives and make it difficult to produce results. Therefore, it is important to ensure that there is clear communication with staff on the objectives of data analytics roadmap.
Data analytics is only powerful when the insights it generates is actionable, this means insights that can be used to tackle issues on hand.
When organisations talk about data science and artificial intelligence, where do you think they should start? What are the must-not-ignore factors? A lot of time we see organisations talking about launching a futuristic technology driven vision but fail to follow it up with an execution plan? What would you like to suggest them?
It is essential for organisations to start with a clear vision of what they want to achieve. They should carefully consider how AI and data analytics can better help the organisation improve its day to day work and achieve its broader goals. This will aid in obtaining buy-in from employees and smoothening the transition to a more data-driven environment. In addition, there must a strong signal from top management to encourage staff to adopt a data-driven mindset.
Adoption of AI and data analytics will change business models and processes within an organisation, affecting the composition and skillsets of its workers. Organisations should also place emphasis on up-skilling employees to encourage all staff to have a stake in the transition.
At the same time, organisations should have a good appreciation of what AI and data analytics can achieve, identify the pain point or problem statement and have a clear roadmap to implement the solutions. Proper governance processes also have to be in place to ensure that organisations are able to obtain data of the necessary quality, before advancing in their data analytics journey.
What are some of the noticeable achievements of MAS and the Government of Singapore when it comes to Data Analytics and Artificial Intelligence?
Singapore has embarked on the journey to become a Smart Nation, to harness the power of technology to support better living, increase productivity, and create new jobs.
At MAS, we believe that a Smart Nation needs an "open API economy" with "connectors" that allow systems to talk to one another, enabling service providers to harness information from multiple sources and produce holistic solutions for customers.
To achieve an "open API economy", MAS has worked with the industry to release the following: a. The Finance-As-a-Service API Playbook, so that banks have a common guide to identify and develop APIs; b. The Financial Industry API Register, so that FinTech start-ups have a one-stop shop to explore open APIs that have been made available. To date, more than 270 open APIs have been made available by the Singapore's financial industry.
MAS also launched the AIDA Grant to support the adoption and integration of AI and data analytics in financial institutions, as well as to encourage financial institutions to address resultant workforce impact through proactive up-skilling, re-skilling and redeployment.
We have seen some of the most popular technology entrepreneurs divided on their view when it comes to AI? What are your views on AI being a threat to humanity?
It is undeniable that AI and data analytics allow organisations to better harness their data and extract better insights. However, as more organisations adopt these techniques to augment or replace human decision-making, the risk of misuse is heightened, whether intentionally or unintentionally. This is a particular concern especially in the case of deep learning models or unsupervised learning, where an algorithm or machine is able to receive inputs, "learn" about how these inputs interact and lend themselves to outcomes, and effect decisions or actions. In most of such cases, the "learning" that the machine does is a black box, leading to decisions or outcomes that could be unexplainable, not transparent and potentially unfair.
MAS has announced that we are bringing an industry group of data analytics and financial sector thought leaders to develop a guide to promote Fairness, Ethics, Accountability and Transparency (FEAT) in the use of AI and data analytics in the Singapore's financial sector. The guide, will comprise a set of key principles, to facilitate financial institutions in taking practical and actionable steps to achieve FEAT in their organisations. The guide will also be applicable across all segments of the financial sector, including both regulated and unregulated entities, and financial service providers such as FinTech firms.
Apart from Singapore, if you had to choose three other cities which you feel are doing an excellent job at promoting the culture of technology, which would they be?
We have seen a shift in landscape where more and more companies are leveraging on technologies to enhance their product and services offerings. There is also a rapid increase of technology start-ups operating in the financial sector. Different cities have adopted different strategies in developing the local FinTech ecosystems. It is hard to shortlist just three. While some cities have started on this journey earlier than others, we could definitely learn from each other. For example, how do we balance regulations such that it does not stifle innovation yet still accord consumers the required protection.
What are the five trends to watch out for in the Fintech sector when it comes to Artificial Intelligence and Data Analytics?
The key trends we envisage for 2018 are: Intensifying the allocation of resources - firms will likely devote more resources and manpower to harness the capabilities of AI and data analytics to aid their decision-making processes and improve their work efficiencies. Jobs will be redesigned around the AI and data analytics ecosystems. Systems will be reconfigured or replaced so that data can flow seamlessly across platforms and/or predictive models etc.
Understand the black box - AI models suffer from the 'black box syndrome' - there is little insight to how to algorithms reach their outcome. This has led to a movement called 'Explainable AI', where machine learning techniques are designed to produce more explainable results, which maintain prediction accuracy. The adoption of cloud-based solutions - there will likely be an accelerated push towards adopting cloud-based solutions.
Blockchains - It has also been identified as having the potential to facilitate certain aspects of the AI implementation. Blockchains can provide a secure environment for big data owners to connect with AI developers. By doing so, complex machine learning algorithm can be developed to help smart devices take advantage of the data available to them.
One of the most important aspects in FinTech is cybersecurity. How do you see AI helping develop robust security systems?
There is potential for AI to augment the current capabilities of existing security technologies and skillsets of security professionals. Existing solutions face limitations when combating cyberattacks, hence AI with its strong computing capabilities to learn, decipher and pull out indicators of compromise will aid in the organisation's defensive abilities. In fact, AI-enabled security technologies are now available and used in endpoint security solutions which can detect sophisticated malware capable of evading traditional, signature-based anti-virus. We foresee that it is just a matter of time when cyber attackers will also use AI for their cyberattacks, such as an exploit toolkit that can self-learn and reconfigure itself during an attack, in the reconnaissance, weaponisation and attack stage. A recent report on the malicious use of AI also quoted several examples, such as an adversarial AI increasing the effectiveness of social engineering attacks by generating realistic and personalised phishing emails, or malicious bots that evade detection by learning and emulating human behaviour.
We have seen that utilising unlabelled data is a challenge for predictive AI. How do you think that can be solved?
Much has been made about the predictive power of AI. But little airtime is invested in the underlying requirements that have to be in place before predictive AI becomes beneficial.
The reality is that models and codes cannot instill intelligence into a software. Intelligence is acquired, though an adequate training sample. Labelled data allows the model to learn and to correct its mistakes so that it can accurately predict the outcomes in a testing dataset. This is a prerequisite in most AI models that are being discussed and used today.
Unfortunately, there is no magic bullet here. However, there have been significant efforts, such as with Generative Adversarial Networks, to automatically generate labelled data to be used for model training.
Last but not the least, what can delegates expect from your presentation on AI and data analytics in Dubai?
Will be touching on the comparison of AI in theory and the evolution of AI in practice, and what it means for the financial industry. I will also be sharing initiatives that MAS is undertaking to facilitate healthy development of data analytics and AI in the financial industry.
MENAFN1904201800490000ID1096747161