Fintech has brought technological changes to many traditional financial sectors and raised new issues for legislation. Based on the research blueprint of artificial intelligence investment consultants, the practical dilemma of this technology is analyzed and studied, solutions are provided from the perspective of legislation, and draft legislative proposals are proposed. Through empirical data, this paper shows the practical dilemma of the development of artificial intelligence investment consulting in mainland China, and analyzes the reasons for the dilemma from the legislative provisions and legislative concepts. Using comparative analysis, this paper examines the legislation of the birthplace and major market regions of AI investment advisors, and provides two legislative paths to solve the dilemma.
I. Introduction
Artificial Intelligence Investment Advisor (hereinafter referred to as RoboAdvisor), also known as Automated Financial Investment Advisor, is a digital financial investment advisor that can automatically establish and manage investment portfolios according to investors' investment preferences. Robo-advisors usually provide online investment management services, and the underlying technology of this service is to use mathematical rules or algorithms to calculate the most efficient path to manage and allocate investors' assets. These algorithms are executed by software and do not require human advisors to relay investment recommendations to investors. On the face of it, human intervention is limited to a minimum in the services provided by robo-advisors.
It is believed that the first year of the emergence of robo-advisors was in 2008, and the entire robo-advisory system was established by several financial start-ups in Silicon Valley. After the robo-advisor enters the mainland market, the technical operation process is mainly divided into five steps: the first step is to collect user data, including basic information, financial status, and risk appetite; Screening and analysis; the third step is to profile the user, label the user information, extract specific user information, and match the data with multiple channels; the fourth step is to make personalized recommendations to the user based on user needs and product risks; the fifth step is to analyze the law of user behavior from the data in the process of using the robo-advisor, and formulate an asset allocation plan in combination with the algorithm.
After the robo-advisory entered the mainland, it was once sought after by all parties, and its share of assets under management once occupied the top five positions in the global robo-advisory market for a long time. However, in recent years, the trend has weakened and the scale of development has shrunk. Today, with the popularity of artificial intelligence, the development dilemma of robo-advisors in the mainland financial market has become an urgent problem to be solved in the mainland's financial empowerment.
2. Dilemma: robo-advisors that are weakening against the trend
With the popularity of ChatGPT sweeping the world at the beginning of this year, the maturity and acceptance of artificial intelligence technology has entered a new era. Before this technology entered the public eye, it was widely used in both the continental and global financial fields. However, contrary to the vigorous development trend, the development of robo-advisors in the mainland financial market has encountered bottlenecks and even a downward trend.
(1) Artificial intelligence in global finance
Since the development of artificial intelligence technology, it has been widely used in the financial field of mainland China. In the field of financial supervision, AI can be applied to capital market supervision, new financial supervision, debt market risk, personal credit anti-fraud, anti-money laundering, etc. In financial institutions, artificial intelligence can be applied to intelligent risk prediction, intelligent marketing, intelligent investment advisory, intelligent investment research, etc. In the field of financial services, it includes credit risk assessment in the banking industry, life cycle value assessment of bank customers, and management of lost customers. From the current practice, there are many common scenarios in the three major financial industries of banking, insurance, and securities, such as customer identification, intelligent risk control, intelligent marketing, intelligent customer service, intelligent compliance, and intelligent operation. Among them, the banking industry has the most extensive scenarios for adopting and implementing AI technology, especially state-owned banks, large joint-stock banks, and large financial groups. The securities and insurance industries have a relatively low level of maturity in the use of AI.
Artificial intelligence (AI) technology is also widely used in the financial sector outside the region. According to information gathered by Forbes in 2020, 54% of financial services institutions have more than 5,000 employees already using AI technology. J.P. Morgan Mercantile Banking, for example, uses key fraud detection applications, which include algorithmic detection of fraudulent behavior patterns. The financial consumer's credit card transaction data is sent to a data center, which processes to determine whether the transaction is suspected of fraud. Goldman Sachs, Bank of America, and Merrill Lynch are all using AI analytics products developed by Kensho.
(2) Robo-advisors that do not increase but decrease
Providing a platform Statista based on professional market and consumer data Data updated in April 2023 shows that from 2017 to 2022, the size of assets under management by robo-advisors increased from $0.19 trillion to $2.45 trillion, the average client using robo-advisors invested in assets under management increased from $4,410 to $6,320, the total revenue generated by robo-advisors for clients increased from $714 million to $9.426 billion, the average revenue per client increased from $320 to $460, and the number of people using robo-advisors increased from 21.99 million to 156.9 million. During this six-year period, the United States and China have always occupied the top two markets in the global robo-advisory market.
However, it is worth noting that the development of China's robo-advisory market is gradually moving towards the opposite trend of the global market. Between 2010 and 2013, robo-advisors were in their infancy in the world, and although the number of companies providing robo-advisors in China grew steadily, they also hovered around 10, and in 2014 and 2015, China's robo-advisory companies ushered in explosive growth, reaching nearly 50 companies at the most. But soon, starting in 2016, the number of companies providing robo-advisory services in China began to decline, and the number of companies in 2018 fell by nearly 60% compared with 2017. In addition to the survival of the fittest caused by competition among enterprises in the robo-advisory market, there are also policy factors. The Guiding Opinions on Regulating the Asset Management Business of Financial Institutions (hereinafter referred to as the "New Regulations"), which came into effect in 2018, restricts the qualifications of robo-advisory business. In 2022, the securities regulatory authorities in many places issued notices on fund investment advice activities, such as pointing out that "some institutions confuse fund investment advisory business with fund distribution business incidental fund investment advice activities". Due to the relevant content of the notice, many financial institutions have taken the initiative to take the robo-advisory business offline. According to the current data and trends, according to Statista's prediction, the size of China's robo-advisory market will fall out of the top five positions in the world from 2023.
3. Attribution: Institutional design restricts technological development
The crux of the reason why robo-advisors have walked out of the opposite trend of the global market in mainland China is that the relevant legislation in mainland China restricts the core functions of robo-advisors to "wealth management on behalf of customers". The reason why the legislation restricts the business scope of the robo-advisor's field is also related to the concept at the beginning of the legislation.
(1) The current legislation restricts core technologies
From the perspective of legislation at the national level, only Article 23 of the New Asset Management Regulations has clear provisions on robo-advisors. Paragraph 1 of this article requires that robo-advisory services shall obtain investment advisory qualifications, and non-financial institutions shall not carry out asset management business beyond the scope or in disguise. Combined with the contents of paragraphs 2 and 3 of this article, at this stage, only financial institutions can use robo-advisors to carry out asset management business. Paragraph 2 is a provision on the use of AI technology by financial institutions to carry out asset management business, and the content of the provisions is consistent with the provisions on non-AI asset management business, such as the review of the suitability of investors, and also reflects the special provisions on AI technology, such as the main parameters of the model and the main logic of asset allocation, and the inherent defects and use risks of AI algorithms are fully reminded. Paragraph 3 is the overall requirement for artificial intelligence algorithms to avoid the adverse consequences of algorithm homogenization on the market.
The China Securities Regulatory Commission (hereinafter referred to as the China Securities Regulatory Commission) has issued the "Interim Provisions on Strengthening the Supervision of the Use of "Stock Recommendation Software" to Engage in Securities Investment Consulting Business (hereinafter referred to as the "Stock Recommendation Software Regulations") on stock recommendation software. Therefore, this provision has certain reference significance. Articles 2 and 4 of the Provisions on Stock Recommendation Software require that the provider of the software shall hold the qualification of securities investment consulting business, that the software shall treat customers fairly, that the same products shall not be sold to customers at different prices, and that the whole process of marketing and service shall be traced.
Robo-advisors are technological upgrades of investment advisors, so the Interim Measures for the Administration of Securities and Futures Investment Consulting and the Interim Provisions on Securities Investment Advisory Business related to investment advisors will also have an indirect impact on the development of robo-advisory business. The former is an administrative regulation, promulgated on December 25, 1997, and Article 24 (1) prohibits investment advisers from engaging in securities and futures trading on behalf of investors, while the latter is a departmental regulation, and Article 31 prohibits investment advisers from making investment recommendations to buy, sell, or hold specific securities. The robo-advisory business model provides investors with personalized recommendations, and even the function of "one-click buy/sell" is easy to contradict these two prohibitions. However, the direction of the development of robo-advisors is to optimize these investment functions. Article 159 (1) of the Securities Law promulgated on December 29, 1998 fixed this rule of the lower law by the superior law.
In fact, the business scope of investment consultants is very extensive, insurance asset management institutions can hire professional service institutions to provide investment advisors for products, bank wealth management subsidiaries also need investment advisers, and even securities and futures operating institutions entrust third-party asset management institutions to provide investment advice when managing products. As long as there is a business need for investment advisors, it means that there will be a place for robo-advisors as well. In addition to the above-mentioned territory, such as Goldman Sachs and Bank of America, the general trend of using artificial intelligence technology for investment analysis is very likely to develop in China, and the mainland has an urgent need to cultivate and expand the artificial intelligence industry and accelerate the overall development of artificial intelligence military and civilian in the future.
From the combing of the existing regulations, it can be found that there are three major problems in the current legislation on robo-advisors in the mainland. First, there is a lack of detailed targeted rules. Although the New Asset Management Regulations require financial institutions to report the main parameters of the model and the main logic of the configuration in response to the characteristics of robo-advisors, and to remind investors of inherent risks, there is a lack of detailed rules on the extent to which the content of reporting and disclosure should be refined, and on whether and to what extent financial institutions can be artificially involved in the process of providing robo-advisory services. Second, the existing regulations restrict the development direction of robo-advisors. As mentioned above, the operation mode of robo-advisors conflicts with the existing prohibitions on investment advisors, and the biggest feature of robo-advisors is that they can replace the labor costs of institutions and investors, automatically match user needs, and balance investment portfolios. This function of robo-advisors represents the future development direction of financial technology, and is also an important path to inclusive finance and common prosperity. Third, there may be inconsistencies between existing provisions. Article 23 of the New Asset Management Regulations stipulates that financial institutions providing robo-advisory services shall "strictly monitor the trading positions, risk limits, types of transactions, price authority, etc. of the intelligent management accounts".
Since 2020, the Standing Committee of the National People's Congress has put the field of artificial intelligence in the annual legislative work plan for three consecutive years, from attaching importance to research in the field of artificial intelligence, accelerating the completion of relevant shortcomings, and accelerating the pace of legislation in this field. The above issues can be gradually resolved in the following legislative projects.
(2) Tracing and reflecting on the current legislative concept
The mainland's investment advisory business system is borrowed from overseas mature markets, but it is not allowed to carry out discretionary account entrustment by licensed investment advisory institutions because they are concerned that investment advisory institutions will use client accounts to manipulate the market or seek personal gain. It can be inferred from this that the original intention of the mainland's top-level design to restrict the business scope of investment consultants was to ensure fairness and justice in the market and protect the investment groups dominated by small and medium-sized retail investors from encountering unfairness due to information asymmetry.
Reflecting on the investment advisory industry itself, the existing prohibitions have led to the slow development of this industry after more than 20 years, and among the registered securities investment consulting institutions, only a small number of companies with strong brokerage backgrounds have developed well, and most of the companies have not been able to expand their scale, and the entire industry is sluggish. There are even individual institutions and practitioners who violate the rules in order to improve their profitability. From the perspective of industry development, the fairness of the current prohibition to industry practitioners is debatable. The financial industry provides a large number of jobs for society, and the benign development of the investment advisory field is bound to release some of them. In the current severe situation of employment pressure, it is necessary to shift from only considering the protection of investors in the past to comprehensively considering the interests of the market as a whole to achieve more comprehensive fairness and justice.
From the perspective of the development of emerging robo-advisory technology, as mentioned above, this technology has broad development prospects and upward trends in the world, because robo-advisory technology reduces the cost of investment consulting and advice provided by financial institutions, so that ordinary investors who do not have huge wealth can enjoy professional advice, and this technology is inclusive finance in the true sense of the word, which has an important contribution to the goal of common prosperity in mainland society. The current regulations limit the technology's core service – managing portfolios directly for users. From the perspective of business practices outside the region, because robo-advisors minimize the human factor and are not affected by emotions when investing, the returns generated by this core service are very stable and considerable. Therefore, it is necessary to change the existing restriction model to benefit more small and medium-sized investors, and to prevent the development of the mainland industry in this area from lagging behind the pace of the world.
IV. Comparison: Extraterritorial Legislative References
(1) United States: Special provisions are made for algorithms
In the United States, where robo-advisors were born, there is no specific legislation for this technology, and robo-advisory services provided by financial institutions are governed by the existing Investment Advisory Act of 1940 and are subject to securities-related laws and regulations. As a result, robo-advisors, like human investment advisers, need to register with the SEC as investment advisers. An investment advisor may provide investment advice to a client and, depending on the client's authorization, may decide to buy or sell securities or other assets. In order to pass legislation as soon as possible, the content of the Investment Advisers Act of 1940 was relatively loose and did not comprehensively stipulate the obligations of investment advisers. It was not until 1962 that the U.S. Supreme Court ruled in the U.S. Securities and Exchange Commission v. Capital Gains Research Corporation that expanded the interpretation of Section 206 of the Investment Advisers Act of 1940, which was later held to establish the rule that investment advisers owe fiduciary duties to investors, and the SEC repeatedly regulated the fiduciary duties of investment advisers. The fiduciary duties of an investment adviser include that the investment advice provided by the investment adviser to the client should be commensurate with the client's financial situation, investment experience and investment objectives, that the investment adviser should ensure that the progress made by the client is the most cost-effective option for the client when recommending a trader to execute the transaction, and that the investment adviser should not have a conflict of interest with the client and should disclose it, if any.
In 2017, the Investment Management Department of the U.S. Securities and Exchange Commission issued guidelines to regulate the characteristics of robo-advisors, including information disclosure, investor investigation, and procedural compliance. Among them, the section on information disclosure is the most characteristic of robo-advisors, that is, the content of the disclosure should explain how the robo-advisor's algorithm manages the investor's account (e.g., how to calculate portfolio recommendations and how to rebalance the investor's portfolio), describe the assumptions and limitations on which the algorithm is based (e.g., the portfolio principle on which the algorithm is based and the flaws behind the principle), and describe the inherent risks of the algorithm (e.g., the algorithm does not take into account market factors when rebalancing the investor's account, or the algorithm performs investment operations more frequently than investors expect, or the algorithm cannot address long-term market changes) (e.g., a description of the circumstances in which the algorithm may be subverted) (e.g., temporary measures taken when market conditions become severe), a description of the circumstances in which a third party may be involved in the algorithm, including any conflicts of interest that may arise from such involvement, details of the fees charged by the robo-advisor to the client and the costs incurred by the client (e.g., fees incurred by the client in using the robo-advisor to regulate mutual funds, or other commissions), an explanation of the extent to which human involvement is in the robo-advisor, a description of how the robo-advisor uses the data collected from the client to generate portfolio recommendations and any associated limitations, and an explanation of how and when the client should update the data in the robo-advisor。
(2) The European Union: the legislative principle of "technology neutrality".
At present, there is no specific legislation for robo-advisors in the EU, and robo-advisors are required to comply with the EU Directive 2014 on Markets and Financial Instruments (hereinafter referred to as MiFID2). The legislative philosophy of MiFID2 is "technology neutral", i.e. the legal rules apply to all types of technology, regardless of the legal rules. The EU allows robo-advisors to rebalance and renew their clients' portfolios after entering into an investment agreement with them, which falls under the practice of "managing portfolios" as defined in Article 4(1)(8) of MiFID2. Therefore, all robo-advisors that provide this service must be approved by the competent authorities of the country in which they are located and be supervised by regulatory authorities. According to Article 23 of MiFID2, robo-advisers should avoid and disclose to clients information about conflicts of interest. According to Article 24 of MiFID2, the services provided by robo-advisors should be honest, fair, professional and in line with the requirements of maximizing the interests of clients. Article 25 of MiFID2 requires robo-advisors to have an accurate assessment of the suitability of investors.
Article 17 of MiFID2 is the only provision in the law that does not adopt the principle of "technology neutrality", which is for algorithmic trading. According to paragraph (1) of this article, investment companies must follow the operational guidelines to prevent erroneous instructions from being issued in algorithmic trading and avoid disrupting the trading market due to systemic reasons. Investment firms must establish effective systems and risk control mechanisms suitable for their own business to ensure that the trading systems they provide are sufficiently resilient and capacitive, comply with appropriate trading restrictions, and fully test and monitor the systems. Paragraph (2) of this article requires investment firms to notify the supervisory authorities of the content of the algorithmic trading in which they are involved. The supervisory authority may require the company to provide relevant documents to demonstrate that it meets the requirements of subparagraph (1) above. The content of the relevant materials should include: a description of the strategic nature of algorithmic trading, details of compliance and risk control in the process of testing the system. The frequency of the provision of relevant materials can be regular or temporary. The supervisory authority has the right to request information about algorithmic trading and the use of the system at any time.
The technical advantage of this EU legislation is that few new technologies are not regulated by law, and the disadvantage is that there is often uncertainty among market participants about how these rules will be applied to new technologies.
5. Pathfinding: China's legislative path for robo-advisors
In order to reverse the weak status quo of robo-advisors, it is necessary to start with the most fundamental revision of legislation, which includes a change from the initial value concept of only protecting the rights and interests of small and medium-sized investors, to balancing the development of the industry and encouraging the development of inclusive finance.
(1) Equitable concept: the transformation and equilibrium of legislative concepts
From the perspective of legislation, it can be seen from the above analysis that the system of restricting investment advisers from acting as sole agents for clients has its own special legislative background, which is the result of the consideration of the top-level designers to protect small and medium-sized retail investors. After more than 20 years of practice, the cultural level of investors and the industry background of investment consultants have changed dramatically. The current prohibitions can no longer fully adapt to the market, and even from the perspective of the market as a whole, the scope of fairness and justice achieved at the beginning of the legislation needs to be expanded.
In addition, in the new round of institutional reform plan of the State Council adopted by the National People's Congress and the National People's Congress in 2023, the relevant measures on the reform of the financial regulatory system are also conducive to the transformation and implementation of the above-mentioned legislative concepts. "After the reform of the mainland's financial regulatory system, the new structure of 'one bureau and one meeting' will reflect unique Chinese characteristics: the conduct supervision with investor and consumer protection as the core and the maintenance of market order as the purpose will be unified under the Financial Supervisory Administration, and the micro-prudential supervision with the sound operation of financial institutions as the core and the prevention of market risks as the purpose will retain the separate business model, and the Financial Supervisory Administration and the China Securities Regulatory Commission will be responsible for the prudential supervision of the non-securities industry and the securities industry respectively." The "mixed-industry" supervision of investor and consumer protection can reduce the problem of regulatory information asymmetry in the financial industry, and can effectively monitor the possible professional problems of investment advisers, so the business scope of the industry can be appropriately liberalized. In particular, a relatively low-level legislative pilot can be carried out for robo-advisors, a new form of investment advisory, so that true inclusive finance can benefit more small and medium-sized investors, and the mainland can also avoid missing out on its priority position in the field of financial technology.
(2) The choice of legislative path with Chinese characteristics
Through the above analysis, it can be seen that it is urgent to revise the legislation to provide development space for robo-advisors. Judging from the current legislative structure, there are two ways to amend legislation. One is to start directly from the high-level legal level and relax the business scope of investment advisers from the top down to allow carte blanches. The other is to continue to make legislative changes only for robo-advisors, and only allow robo-advisors to accept discretionary mandates from clients.
According to the first approach, the provision prohibiting investment advisers from "engaging in securities investment on behalf of the client" in Article 161, Paragraph 1, Subparagraph (1) of the current Securities Law should be deleted. Correspondingly, similar prohibitive provisions in the Interim Measures for the Administration of Securities and Futures Investment Consulting and the Interim Provisions on Securities Investment Advisory Business also need to be deleted. The advantage of amending the legislation in this way is that it can completely solve the problem of the narrow business scope that has plagued investment advisers for a long time, and at the same time, it can also solve the problem of business restrictions of robo-advisors. This approach is similar to the EU's "technology neutrality" principle, which regulates new technologies with a universal rule. Since it is impossible to circumvent the securities law as a superior law to expand the scope of investment advisory business, the disadvantage of choosing this path is that it will take a relatively long time to push for the law to be amended, and there is uncertainty about whether it will eventually be passed. Because after all, it is an opening restriction for the entire industry, and the game of the interests of all parties in the early stage and the consideration of supervision in the later stage are all factors that legislators need to consider.
Another approach is to liberalize the business scope of discretionary investment advisors only, similar to the current legislative strategy of the new asset management regulations. However, the New Asset Management Rules Article 23 currently has two problems: the first is that the logic of the rules is not clear enough, and the first half of the first paragraph requires that institutions providing robo-advisors should obtain investment advisory qualifications, which is combined with the restrictions on investment advisers in the Securities Law and other laws, and it can be seen that it is aimed at robo-advisors who only provide investment advice, while the latter part of the paragraph prohibits non-financial institutions from carrying out asset management business with the help of robo-advisors, and the second half of the article mentions that financial institutions use robo-advisors to carry out asset management business, which is to allow this part of robo-advisors to provide discretionary business。 There is a conflict between the provisions of the two parts as to the positioning and business scope of robo-advisors. The second problem is that the provisions of Article 23 mainly focus on the regulation of information disclosure obligations, including algorithms, which are generally general and need to be further refined.
To solve the first problem, it needs to be combined with the first path scheme mentioned above. If the future legislation chooses the first path and fully liberalizes the business scope of investment advisers, then even if the original provisions of the first half of paragraph 2 of Article 23 of the New Asset Management Regulations are retained, there will be no logical conflict in that provision. However, if the second route is followed, only the business scope of robo-advisors will be liberalized, then this part of the provisions needs to be deleted or re-enacted, and robo-advisors will be explicitly allowed to carry out discretionary business. In addition, one of the issues that need to be considered in the top-level design is whether to open up the scope of robo-advisors that can provide discretionary services, or whether they are still limited to financial institutions. If it is still limited to financial institutions, then it is only necessary to expand on the issues discussed below on the basis of the New Regulations. If there is a need to expand the scope of providers, then it will be necessary to re-establish departmental regulations and other hierarchical regulations for robo-advisors. On the one hand, the current regulatory division of labor and means mentioned above are sufficient to carry out comprehensive management of such technologies, and on the other hand, more qualified enterprises can participate in the R&D and supply of intelligence and investment advisors, so that the competition in this market will be more sufficient, which is conducive to providing more investment channels for small and medium-sized investors, and is also conducive to the development of different algorithms, so as to avoid the risk of "herd effect" in the market due to the homogenization of algorithms.
As for the second question, the regulatory provisions of robo-advisors should provide detailed provisions on the characteristics that distinguish robo-advisors from ordinary human investment advisers, including the fiduciary obligations that robo-advisers should bear to users and the details of algorithmic disclosure. First of all, with regard to the fiduciary obligations of robo-advisory operators, there are already relevant provisions on investment advisory business and asset management business under the existing legal and regulatory framework. Article 136 of the Securities Law, Articles 5, 12, 13, 19 and 26 of the Interim Provisions on Securities Investment Consulting Business, Articles 3 and 28 of the Trust Law, and Article 3 of the Interim Provisions on Strengthening the Supervision of the Use of "Stock Recommendation Software" to Engage in Securities Investment Consulting Business, all of which regulate the duty of loyalty of users, require institutions to warn customers of risks, prohibit the promise of investment returns, and prohibit conflicts of interest. Articles 4, 15, 16 and 19 of the Interim Provisions on Securities Investment Advisory Business, Article 20 of the Interim Measures for the Administration of Securities and Futures Investment Consulting, and Articles 8 and 25 of the New Regulations on Asset Management all stipulate the duty of care that institutions should undertake, requiring them to fulfill their suitability obligations, fully collect relevant information, and undertake the best execution obligations to ensure the best interests of clients.
Robo-advisory operators only need to directly comply with and enforce the above-mentioned existing regulations. In addition to the moral hazard brought by this new technology, which is already subject to the existing fiduciary duty provisions, its unique operational characteristics need to be regulated by new rules, the most important of which is the restriction on conflicts of interest. Because robo-advisors will have a different conflict of interest between the operator and the program designer. "For example, robo-advisory company A invites technology company B, which is proficient in algorithm architecture, to design core algorithms due to its lack of professional technology, and if B accepts the benefits of C and agrees to give priority to the financial products it sells or its company's stocks, then B can manipulate the algorithm based on its technical advantages, so that C's products can always be evaluated for advantages and be widely recommended to customers." Therefore, robo-advisors should be provided with provisions to address such problems, requiring users to disclose information about program designers and potential conflicts of interest. In addition, algorithms are the core of robo-advisors, so the disclosure and supervision of algorithms is an important part of regulating robo-advisors, and should be as detailed as possible to facilitate operators' reference and implementation.
epilogue
The current "New Regulations" not only contradict the legislative logic, but also limit the scope of operators that can provide core functions. These issues need to be addressed through the top-level design of legislation.
Annex I:
Regarding the use of "artificial intelligence investment advisors" to provide securities investment consulting and asset management services
Interim Provisions on Supervision (Proposed Draft)
These Provisions are formulated to further regulate the use of "artificial intelligence investment advisors" by institutions to provide securities investment consulting and asset management services, and to protect the lawful rights and interests of investors.
1. "Artificial intelligence investment advisor" as used in these Provisions refers to applications, software products, software tools or terminal equipment with one or more of the following securities investment consulting and asset management service functions:
(1) Provide opinions on investment analysis, forecasting, and selection involving specific securities investment varieties;
(2) Accept the entrustment of investors to invest in and manage the entrusted investors' property.
II. Where "AI investment advisors" are sold or provided to investors, and economic benefits are directly or indirectly obtained, business qualifications shall be obtained with the permission of the regulatory departments.
Without obtaining business qualifications, no institution or individual may use "artificial intelligence investment advisors" to provide investment consulting business or asset management services.
3. Institutions using "AI investment advisors" to engage in securities investment consulting and asset management business shall follow the principles of objectivity, fairness, honesty and credibility, and shall not mislead or defraud users, and shall not harm the interests of users.
4. When providing "AI investment advisors" to users, institutions shall comply with the following regulatory requirements:
(1) Publicize information on the company's business premises, the company's website, and the website of industry associations, including but not limited to: the company's name, address, contact information, complaint telephone number, business license number, employee's name and registration code;
(2) Follow the principle of user suitability, formulate systems and processes for understanding users, classify and grade the services provided by "AI investment consultants", and disclose the characteristics and risks of services to users, and sell appropriate products to appropriate users;
(3) Objectively explain the functions of software tools and terminal equipment, and must not conduct false, untruthful, or misleading publicity about their functions;
(4) Revealing the inherent defects and use risks of software tools and terminal equipment, and must not conceal or have major omissions;
(5) Treat users fairly, and must not sell the same product to different users at different prices by inducing users to upgrade and pay;
(6) Establish and improve the internal management system to achieve objective, complete and comprehensive traces of the service process, and archive and manage the records of the traces; the retention period of relevant business files shall not be less than 5 years from the date of termination of the relevant agreement;
(7) Fully disclose information and potential conflicts of interest of program providers who design, develop, and maintain "AI investment advisors";
(8) Shall not promise or guarantee investment returns to users in any way;
(9) Regularly explain to the regulatory authorities how the algorithms of the "AI Investment Advisors" manage investors' accounts, revealing the inherent risks of the algorithms and the circumstances under which the algorithms may be subverted;
(10) Regularly disclose to the regulatory authorities the extent of human involvement in "AI investment advisers";
(xi) regularly describe to regulatory authorities how the robo-adviser uses the data collected from the user to generate portfolio recommendations and any related restrictions;
(12) Periodically explain in detail to users the fees charged by the "AI Investment Advisor" and the costs that the users need to bear.
5. Institutions providing investment consulting services through "AI investment advisors" shall strengthen investor education and protection of users' rights and interests in each business link, and actively inform users of the company's business qualifications and contact information.
VI. Where institutions and their staffs violate relevant laws and regulations and these Provisions in the course of using "AI Investment Advisors" to provide investment consulting and asset management services, the relevant departments shall take regulatory measures or impose penalties in accordance with law;
VII. Where institutions and individuals that have not obtained business qualifications use "AI investment consultants" to engage in illegal investment consulting and asset management activities, the regulatory authorities shall cooperate with local governments, public security organs, judicial organs, and so forth in accordance with the provisions of laws and regulations to investigate and deal with them in accordance with law;
8. These Provisions shall come into force on the date of ****************