To what extent are you aware of artificial intelligence already being used in the financial sector in your jurisdiction, and do you think regulation will impede or encourage its further use?
Fintech (2nd edition)
Artificial intelligence (AI) technology is gaining wider adoption across the financial services industry as businesses learn of the benefits that AI offers to a whole host of processes. Fuelled by the deluge of data created over the past decade and the meteoritic rise of computing power thanks to cloud computing, the U.S. financial industry is adopting AI at an accelerated pace, revealing a sharp increase in the number of uses of AI technologies in the financial sector. For instance, proprietary trading firms and hedge funds have long used algorithms that trade on an automated basis. So-called “automated trading” or “high-frequency trading” has come under the scrutiny of the SEC and the CFTC, particularly following the “flash crash” in 2010. Both agencies have considered new regulations to address these concerns and have brought enforcement actions against trading firms to police the use of such algorithms to manipulate securities and commodities markets.
Service providers have designed both retail and institutional trading software products that automate the selection of investments for the trader’s portfolio and trade or provide trade signals based on an AI algorithm (such products oftentimes referred to as “robo-advisors”). These software providers may be regulated as investment advisers or broker-dealers by the SEC or as commodity trading advisors, introducing brokers, or futures commission merchants by the CFTC. Moreover, robo-advisors raise novel consumer protection issues for state and federal agencies.
Financial institutions and software companies are also incorporating AI into applications to help consumers manage their finances. These applications may make suggestions to users about how to allocate money or automatically move funds on the user’s behalf. These businesses must consider banking, money services, and consumer protection laws.
Certain initiatives using AI are being developed, especially in the investment industry (robo-advisory, etc.), credit (especially credit scoring), and regtechs (automated KYC, etc.). Chatbots are also getting more and more common in the finance industry.
At this stage, regulations dedicated to AI processes are rather seldom. One of the most important exceptions can be found in the GDPR which has put in place specific requirements for automated decision-making processes.
We believe that specific AI requirements are likely to increase in the future. They will probably create new safeguards which will on the one hand restrict the use of AI but on the other hand will also offer a better legal framework and increase legal certainty in the field.
Artificial intelligence (AI) is certainly an area of growth in the financial sector in Bermuda. We are seeing start-ups developing AI for things like robo-advisory services in the investment and funds sector as well as cyber-underwriting in the insurance sector.
The regulatory regime and tech-friendly environment in Bermuda encourages AI’s further use and development within the jurisdiction.
The Brazilian Central Bank encourages the use of artificial intelligence in the financial sector. The lending, compliance and Know Your Client (KYC) sectors are already implementing it.
There is, however, some grey area regarding the applicable laws. So far, the technology is manly regulated by LGPD, which is causing some debate towards the sufficiency of such normative to protect such sensitive data. Up to this point, it is likely that the financial institutions themselves will be in charge of regulating the issue (through what Central Bank called “self-regulation”).
Therefore, even though there might be some issues since the topic is not yet regulated to the extent of allowing provisions with a higher degree of accuracy, the scenario is, overall, optimistic.
Banks and Retailers in Chile started to prove AI for virtual customer assistants (chatbots) to improve user experience (UX), followed by the automation of their internal processes, such as anti-money laundering (AML) protocols. There is also the existence of robo-advisors which are computer-based algorithms that recommend certain investment portfolios according to the user’s preferences, like age, risk-aversion, expected cash returns, length of the investment, amongst others. An example of the latter is Fintual which uses AI to recommend to an investor one of their diversified portfolios, while also serving as an investment platform. Another example is FOL which is an AI platform to invest in mutual funds.
Currently, there is no legislation that impedes or discourages this industry. It has been favorable to innovation that regulation is often behind technology. The only piece of prospective regulation is the White Paper. President Piñera instructed the Ministry of Science to come up with an Artificial Intelligence Policy by April 2020. In our opinion, regulation will be neutral – neither impeding nor encouraging – the use of IA in Chile.
The artificial intelligence (AI) market in China is seeing strong growth, particularly in the following areas within the fintech sector:
(i) Speech recognition and natural language processing applications, such as smart customer service, deep-dive behaviour analysis and voice data mining; and
(ii) Machine learning, neural network applications and knowledge maps, each for applications such as wealth management, fraud detection, intelligent risk control and robo-advising.
Chinese regulators are neither encouraging nor impeding AI technology, though the onerous licencing and other requirements for financial advisory services related to publicly traded funds and securities necessarily entail considerable regulation of AI.
Certain ‘intelligent advisers’ in the private investment sector (e.g., asset managers exclusively providing services to qualified investors) are subject to a different set of regulatory requirements regarding the use of AI, which have been comparatively lax but have recently been subject to increased regulations that will impact China’s Fintech sector throughout 2019.
As the latest general guideline for the asset management industry, the Guiding Opinions for Regulating Asset Management Business of Financial Institutions, jointly promulgated by the PBOC, China Banking and Insurance Regulatory Commission, China Securities Regulatory Commission and SAFE on 27 April 2018) further reiterate that any institution which wishes to use AI technology to engage in investment advisory services must obtain an investment advisor licence. Non-financial institutions are prohibited from engaging in asset management activities under the alternative name of ‘intelligent investment advisers’ if their registered business scope does not include ‘asset management’.
Because China has increasingly been opening up its asset management market (including its market of securities investment fund managers) and has allowed seasoned foreign investment management firms such as JP Morgan, BlackRock and Fidelity to set up wholly foreign-owned private fund management services onshore, it is foreseeable that these foreign giants could be able to further tap into the Chinese private intelligent advisory services market by targeting qualified investors in China.
There are a number of fintech suppliers who are using artificial intelligence actively in enhancing existing financial processes. One such is Eigen Technologies, which is using natural language processing to pull specific data fields out of large amounts of legacy documents in order to help financial institutions to get digital control of the data that they hold in other formats, in a fraction of the time that it would take humans to carry out the same task. There are other examples of regtechs that are using machine learning to extract and package up regulatory information, such as Cube Global.
Other companies are using AI to spot behavioural patterns and anomalies in those patterns – one of AI’s strong suits. These include:
- payments authentication solution Cybertonica;
- email security firm Tessian; and
- Nasdaq Buy-Side Compliance (formerly Sybenetix), which is used by asset managers and hedge funds to spot anomalies and suspicious activity in traders’ trading patterns.
Others use machine-learning to spot patterns in order to make predictions. These include:
- cashflow prediction engines Fluidly and Fractal Labs; and
- insurance pricing and risk engine Cytora.
There is limited regulation in this area at the moment, the main regulation being those parts of GDPR that touch upon the sorts of data processing that are often involved in machine-learning. These will include, in particular, obligations in relation to profiling and automated decision making (see the answer to question 5 above). A House of Lords Select Committee has been established for the purpose of considering the economic, ethical and social implications of advances in AI, and this has focused to some extent on the concepts of ethical uses of broad data sets, accountability, and the mandated sharing of data (similar to open banking) in order to prevent “data monopolies”. It remains to be seen to what extent this thinking will translate into legislation. However, our view is that if any future regulation is technically realistic and flexible, whilst encouraging transparency and accountability, it will be of benefit to consumer trust in its use in the long term, which will in turn lead to widespread adoption.
AI is not heavily used in the Colombian financial sector and is limited to chatbots for bank-client relations. There has been for some years now limited use of trading algorithms in government bond trading which is the most liquid in the country. Regulation may impede the use of AI in several industries as an astringent interpretation of professional responsibility delegation has been in place for a long time. Some advances in securities markets advisory have been adopted allowing for tech-based advisory but are yet to be a common place in the industry.
The use of artificial intelligence in the financial sector is well-established. We do not believe UAE legislation impedes the use of artificial intelligence in the UAE. However, such use would need to satisfy UAE security requirements which are in line with international standards. The Ministry of Finance has on its website included, as a critical activation point to build the UAE’s artificial intelligence capabilities, the issuance of government law and appointment of a UAE Artificial Intelligence Advisory Board to ensure the proper use of artificial intelligence.
Banks use artificial intelligence for a variety of reasons, including the automation of processes, interaction with customers, and building intelligent and real-time lending models. For instance, Emirates NBD has launched the chatbot Eva, which helps customers cut through time consuming layers and access the information they want immediately. Also, First Abu Dhabi Bank (FAB) launched a portal powered by artificial intelligence and machine learning tools to deliver advanced analytics to merchants.
The DIFC has, through its innovation testing program, approved the first robo-wealth advisor in the Middle East, Sarwa. The Dubai Electricity and Water Authority has also launched its robot, Rammas, which listens to customers’ concerns and knows their transaction history.
Artificial intelligence (AI) is already being used in financial and investment services, insurance services, etc. in Taiwan. In particular, compared to expensive financial advisors, robo-advisors are expected to replace humans in some roles and offer low-cost alternatives to manage people’s money. On August 10, 2017, the FSC introduced new deregulatory measures to facilitate the development of robo-advisory services by securities investment consulting enterprises for asset management purpose (i.e., automated rebalancing transactions for the client). However, because of the current regulatory restrictions, only clients with at least NT$5,000,000 to invest would qualify for being accepted for the foregoing automated rebalancing transactions and they would also be subject to predetermined target portfolio weights. Generally speaking, it is necessary for Taiwan to loosen up the relevant laws and regulations so as to encourage further use of AI in the financial sector.
Artificial intelligence (AI) is already used in robo advisory as well as in various compliance projects (e.g. AML, on-boarding of clients) and in document reviews. Regulation will not impede its use; rather, the approach to make regulation technology neutral will encourage its use.
Nets and KMPG have co-developed an AI-powered (Machine learning based) payment fraud prevention solution called Nets Fraud Ensemble. Which have reduced fraud transactions by 25 % during the first month of operation.
The regulation of AI is still in its early form. However, it is an area of focus both in Denmark and on an EU-level. In April the European Commission published, its guidelines for working with AI. In addition, a regulation of AI from EU is in the pipeline within the next couple of years.
No one doubts that artificial intelligence is a powerful tool for the financial sector but not only. However, although it is a trend topic and there is a lot of talk about it, in accordance to experts’ opinions this technology is still very much underway.
The main application of artificial intelligence in financial services is machine learning, specifically applied to credit scores determination. Through the use of machine learning, banks are able to determine the likelihood of a loan being repaid in time or the chances of default, allowing a much deeper analysis into the client’s credit behaviours.
Using artificial intelligence techniques in the provision of financial services can heighten efficiency, reduce costs, enhance quality, raise customer satisfaction levels and boost financial inclusion.
The Bank of Spain issued in March 2019 an Economic Bulletin on “Artificial intelligence in financial services”, referring several positive changes brought possible by artificial intelligence, such as greater automation and analytical capacity.
However, artificial intelligence also poses a series of risks and limitations that must be known and managed so as to be able to correctly extract all their potential.
Bank of Spain considers that cooperation with other authorities, both at a national and a global level, is particularly important in this respect, given that the use of artificial intelligence crosses both national and sectoral boundaries.
According to the results of a recent BaFin consultation, main areas of AI use in the financial sector so far are the analysis of financial and alternative data for trading; product and services pricing; risk modelling; anti-money laundering; and the emergence of platform-based business models. Co-operations between existing businesses and new specialised big data providers/BigTechs are on the rise, resulting in an increased use of outsourcing systems/APIs and a growing fragmentation of value chains.
The existing financial market regulatory framework aims to be principle-based and technology-neutral and should therefore by definition not hamper the use of AI. However, certain AI-related aspects are perceived as regulatory challenges, such as a lack of explainability and traceability of AI-based decisions. BaFin states it will not accept any risk assessment models presented as an unexplainable black box. Further, it discusses new regulatory approaches such as the appointment of algorithm officers (similar to data protection officers), the establishment of data ethics commissions, code review processes, simulation and penetration tests and reviewing sample profiles.
Presently, artificial intelligence technology such as robo-advisor and chat-bot is being utilized in capital markets industry.
The Korean government has designated AI as its main focus business and has announced that it would not spare any support in relation to the development and commercialization of AI. There are, however, claims that laws and regulations relating to personal information are hampering the development of AI—this is because development of AI through the use of deep-learning technology requires utilization of a large volume of big data, which can be restricted by the regulations relating to the use of personal information.
To our knowledge AI is not used in any significant manner in the financial sector. However, AI based solutions have been, and are being, developed in Iceland such as Authenteq and Lucinity. Both are aimed at anti-money laundering, Authenteq in the KYC process and Lucinity to detect suspicious behavior patterns.
AI is a central technology for all sectors and industries, and as a result its testing and use is growing worldwide, including in Portugal. Indeed, AI (together with Big Data) is already being used in Portugal by banks and fintech companies, e.g., in customer service, in the analysis and granting of credit and even in regulatory compliance. AI brings several legal and regulatory challenges, from warranty, liability, transparency, among others. Regulation of AI is being thoroughly discussed at the international level, including in the EU, and such discussions also cover the strategic approach to adopt in this field.
On 8 April 2019, the European Commission-led High-Level Expert Group on AI presented Ethics Guidelines for Trustworthy Artificial Intelligence. According to the Guidelines, trustworthy AI should be: (1) lawful - respecting all applicable laws and regulations; (2) ethical - respecting ethical principles and values; (3) robust - both from a technical perspective while taking into account its social environment. The Guidelines put forward a set of 7 key requirements – which are duly developed in the Guidelines – that AI systems should meet in order to be deemed trustworthy. These include:
- Human agency and oversight;
- Technical Robustness and safety;
- Privacy and data governance;
- Diversity, non-discrimination and fairness;
- Societal and environmental well-being;
Additionally, the Portuguese government has also recently approved its National Strategy for Artificial Intelligence – “AI Portugal 2030”, which sets out the main general objectives to reach by 2030, including:
- Added Economic Growth: the added value brought by AI technologies to the economic growth should be significant;
- Scientific Excellence: improve the front-line position in fundamental and applied AI research of the Portuguese Academia (universities, polytechnic schools and research institutions) measured in terms of publication impact, international leaderships, and international collaborations;
- Human Development: Increase dramatically the qualifications of the labour force, in particular technological qualifications, while promoting inclusion and awareness at all levels of education.
The National Strategy does not specifically address, however, the challenges AI will bring to the financial sector. Nonetheless, it is reasonable to assume that the increasing number of public guidelines and recommendations (both at national and international level) will foster investment in and the development of new AI-driven technologies and businesses.
In this scope, it is important to make sure that any potential AI regulation does not impede or hinder the development and use of AI, and an approach of regulating without a full understanding of the evolution of this technology shall be avoided. .
Artificial intelligence (AI) based tools and utilities have witnessed increased use cases in the financial sector in India, with financial services witnessing a shift from personal customer interactions to automated processes. The primary areas where AI has made strides in the financial services in India relate to product matching in accordance with the requirements and profile of customers, fraud detection, assistive chatbots and personal finance management.
The NITI Aayog has recently proposed an institutional framework for AI which requires funding of INR 75,000,000,000 (Rupees seventy five billion) for creation of cloud computing platform called AIRAWAT. Reports suggest that the Indian government is looking to invest INR 75,000,000,000 (Rupees seventy five billion) initially over a three-year period and set up a high-level taskforce to oversee roll-out and implementation of AI. The Government of India has also proposed setting up a National Centre for Artificial Intelligence to encourage AI based technology including in the FinTech sector.
That being said, the key challenge in connection with use of AI in financial sector in India is with respect to data privacy and protection. Use of AI utilities involves access to sensitive customer data including inter alia financial information, credit history, spending patterns, etc. So long as financial service providers are able to comply with the changing regulatory landscape pertaining to data privacy in India, regulation is unlikely to impede further use of AI in financial services. However, financial service providers will need to re-think their systems and operations so as to ensure a smooth transition to the proposed data protection framework in India and to deploy AI for enhancement of customer experience in banking and payments space.
Artificial Intelligence (AI) started to be used in the financial sector 3 years ago. However, the level of development of this technology is still incipient in the country. Banks begun to prove AI for virtual customer assistants (chatbots) to improve user experience (UX), followed by the automation of their internal processes, such as anti-money laundering (AML) controls. It is expected that in the near future AI would be used as an alternative credit scoring for the unbanked.
In our opinion, regulation will be neutral -neither impeding nor encouraging- the use of IA in Peru.
The Israeli financial sector has begun incorporating artificial intelligence and machine-learning with a wide variety of start-ups utilizing artificial intelligence to provide solutions to the financial sector. Israeli banks are not standing idly by and have shown both an interest to invest in and utilize artificial intelligence. While the Israeli regulator has not given clear instructions as to the boundaries of the use of artificial intelligence in the fintech sector, the Bank of Israel is to issue its first banking license for more than 40 years. The license will be issued to a new venture currently currently called The Digital Bank (which is represented by this firm). Other entities in Israel such as Personetics are using AI to develop technology for use by the banking sector for personalized banking services.
The Dutch financial regulators have published initial guidelines relating to the use of AI and self-learning algorithms in the financial sector. For example, the Netherlands Authority for the Financial Markets published guidelines on the duty of care involved in semi-automated asset management and its views on roboadvice (www.afm.nl/en/nieuws/2018/mrt/doorontwikkeling-roboadvies). The Dutch Central Bank (DNB) also recently published guidelines for the use of AI (www.dnb.nl/en/news/news-and-archive/DNBulletin2019/dnb385020.jsp). The acronym of these DNB guidelines is ‘SAFEST’, which hints at DNB's main message. The guidelines urge financial undertakings to use AI responsibly. AI applications in the financial sector should be Sound; someone must be Accountable; the outcome of AI should be Fair and Ethical; only sufficiently Skilled people should be involved in developing AI applications; and the use of AI should be Transparent and explainable. Responsible use of AI is key to prevent incidents which could have a substantial impact on financial stability.
While the use of artificial intelligence in the financial sector still appears to be limited, we can find several players in this sector attempting to utilize it in their business. For example, certain asset managers recently launched mutual funds that use artificial intelligence to make automated investment decisions. Certain banks announced that they launched new loan programs utilizing artificial intelligence as an automated loan screening tool. Certain insurance companies are also attempting to utilize artificial intelligence to handle insurance claims and examine payments of such claims.
In general, the national government of Japan shows a proactive attitude towards the use of artificial intelligence. In the financial sector, the Financial Services Agency supported the testing of a project whereby an IT vendor and financial institutions attempted to have artificial intelligence undertake primary screening of customers' voices and extract potential compliance breaches and customers' complaints based on them. It recently announced the successful completion of the project with the statement that the use of artificial intelligence would be feasible for this kind of screening process.
There is limited uptake of AI and machine learning in Jersey. We are aware of automation initiatives and developments in fintech, legal tech and regtech which will likely embed themselves in to the Jersey financial services industry in coming years. Factors affecting its further use in Jersey are likely to include:
- the uptake in other jurisdictions (such as the UK) affecting market practice in Jersey, particularly if adopted by large multinational groups such as banks or accounting practices;
- market pressures encouraging the adoption in Jersey to enable Jersey to stay competitive; and
- regulation and understanding of the risks associated with the technology and the management of those risks.
Regulation would impact uptake but it would depend on the terms of the regulation as to whether this would impede or encourage use.
The University of Liechtenstein and Kimera Systems announced a partnership regarding an AI research project in 2019.
Based on the generally positive attitude of the government and the FMA toward technical progress in Liechtenstein and its opportunities for the Liechtenstein financial centre, it cannot be assumed that future regulations will be restrictive. For example, the Deputy Head of Government and Minister of Economic Affairs recently officially opened the Digital Summit on the topic of artificial intelligence.
Mexican financial institutions are not only at the forefront of the implementation of artificial intelligence, but a significant percentage of them are already in a process of digital transformation in order to streamline their processes. Other than these private developments, back in 2016, the RegTech for Regulators Accelerator “partnered with (…) authorities in (…) Mexico to develop tools and techniques for better market supervision and policy analysis” (Gurung and Perlman, 2018), an effort which led to the creation of an access-controlled data storage platform of the CNBV that automatically validates, analyses and reports data submitted by financial institutions in connection with AML requirements. We think regulation, as is (no specific AI law exists currently), will foster AI’s further use in the near future.
In Luxembourg, some private actors developed digital tools in relation to, robo-advice and chatbots. In this respect, the CSSF published its White Paper "Artificial intelligence: Opportunities, risks and recommendations for the financial sector" in December 2018.
These initiatives are, currently, developing at a rapid pace. In general, there is an increasing interest of financial professional for digitalization projects in Luxembourg and it is to be expected that artificial intelligence will be more and more present in the Luxembourg banking sector.
Up until now Artificial Intelligence (‘AI’) is not yet regulated in Malta, however the Maltese Government has published a high level policy document putting forward the vision and policy considerations in the development of national AI strategy. Albeit such a document is still in its consultation phase, it is paving the way towards establishing Malta as a model nation in AI policy. In light of this, the Government of Malta has commissioned the Malta AI Taskforce to draft the AI National Strategy. This taskforce has already launched a consultation on the proposed ethical AI framework document to seek the public’s feedback on guiding AI principles and trustworthy AI requirements as set out in the mentioned document.
More banks are planning to replace their front-office employees with chatbots and conversational user interfaces to serve clients and enhance customer experience, according to multinational law firm Baker McKenzie’s latest report. Similarly, in Malaysia, banks have taken steps in implementing Artificial Intelligence in an attempt to assist their everyday functions.
These banking services are being transformed from human interaction to virtual assistants. Making use of AI, natural language and predictive analytics, banks in Malaysia have implemented chat assistants (Chatbots), which are Artificial Intelligence-driven virtual assistants which help their clients in day to day queries. These chatbots are able to search for past transactions, provide information about financial services and help arrange meetings with financial advisors.
An example would be HSBC Bank launching a Chatbot named Amy for its corporate banking clients. Amy understands both English and simplified Chinese, and can provide prompt response to banking queries 24X7.
Malaysian banks have their own success in chatbots. CIMB Bank launched Enhanced Virtual Assistant (EVA), and since its inception in Dec 2016, EVA has generated around 130,000 downloads and 300,000 transactions. RHB Bank and RinggitPlus, meanwhile, jointly launched My RHB Easy, a Chatbot which gives users the flexibility of applying for a personal loan 24X7, without the manual completion and submission of application forms. Chatbots are not the only AI enabled functionality being used by banks. Financial services providers are also heaping on other revolutionised technological innovation to streamline their business model. Biometrics such as voice, facial, iris and fingerprint are becoming an integral part of authenticating individual customers which allows faster approval and overcoming the challenges of manual signature based verification.
In terms of placing regulations on the use of Artificial Intelligence in the financial sector, one must look at the wider picture. While all of the above looks like a fairy tale for the industry, there is also a dark side to it.
Increasing use of AI will also lead to the extinction of human bankers. Gone will be the days when customers will be greeted by a bright smile when they walk into a bank. Automation is leading to disruption in banking jobs. Malaysian banking jobs will not be immune to automation. The Malaysian Employers Federation has predicted that 50,000 Malaysians are expected to lose their jobs in 2018, the bulk of whom will be from manufacturing, insurance and banking. Disruption in the banking sector is also due to pressure created by the rise of financial technology (FinTech) startups, which can provide innovative financial solutions faster at much-reduced cost.
Most financial services providers are jumping on the AI bandwagon without a clear thought strategy of AI implementation, leading to their dependence on robotics and thus losing their human intelligence. This suggests that there is a mismatch between providers and users of the financial services when it comes to the use of AI enabled technology. Clearly, AI has emerged as both boon and bane for the financial services industry. At one end of the spectrum it is expected to streamline operations by providing efficiency and reliability, while on the flip side, it will disrupt the role of humans in financial advisory.
It all goes back to the role of the policymakers and regulators to find a balance between the Rise of the Machine and the Death of the Humanity. Hence why it is in my view that regulation of the use of Artificial Intelligence in the financial sector is indeed important in helping find that much required balance.
It is reported that AI has been deployed in the financial sector in a number of ways to reduce the financial sector’s exposure to risk, although AI technology is still in its nascent stage.
One reported example is that AI has been employed by banks to improve loan underwriting and fraud detection through machine learning, thereby reducing financial and compliance risks associated with the provision of financial services.
While there is no regulation specifically targeting the implementation of AI, given MAS’ stance that regulation should not “front-run” innovation, any future regulation of AI would likely be facilitative of the adoption of AI technologies, and would also manage the risks that come with the usage of AI. In doing so, there could be increased public confidence in AI, leading to further adoption of AI technology.