Author: Lawyer Shao Shiwei
With the rapid development of the digital economy, artificial intelligence is undoubtedly one of the most innovative and promising fields today. Many AI entrepreneurs and programmers are diving into it with enthusiasm and technical dreams. However, seemingly innovative business models may also harbor many unnoticed legal risks.
This case will deeply analyze the legal risks that AI entrepreneurs, programmers, and technical teams may face when engaging in financial technology, quantitative trading, etc., through Shanghai's first case of using an 'AI stock trading robot' for illegal stock recommendations, and will provide compliance suggestions.
1. Shanghai's first AI stock trading software illegal business case: Company controller sentenced to seven years and nine months
Case overview:
S Company, without approval from regulatory authorities, operated the online platform 'Xindong Quantification', promoting S Company's 'interval arbitrage', 'DIY stock trading robots', and other stock trading products, providing clients with individual stock buying and selling timing, recommending specific stocks, and other securities consulting services. The total profit amounted to over 30 million yuan[i].
After two levels of court hearings, ultimately, S Company's actual controller Zhong was sentenced to seven years and nine months.
2. Is selling AI stock trading quantitative trading software illegal?
In this case, the technology company's boss Zhong believes[ii]:
DIY stock trading robots developed by oneself are based on plans set by customers, helping them filter and buy the stocks they want. In Zhong's view, the purchasing strategies are all formulated by the customers, and the software merely plays a role in data analysis, without recommending stocks, thus the company's operations do not require corresponding qualifications.
However, this view was not adopted by the court. Many AI industry practitioners, including programmers, may also hold similar views:
AI stock trading software is merely an information screening tool, not investment advice. The software automatically collects and organizes publicly available market data (such as capital flow, trading volume, etc.), helping users improve their information processing efficiency. Ultimately, the decision-making power is entirely in the hands of the user, which fundamentally distinguishes it from providing specific buy/sell recommendations or promises of returns, as in 'stock recommendations'. Therefore, they believe such software belongs to 'neutral information technology services' and does not constitute 'illegal securities investment consulting business' as stipulated in Article 225 of the Criminal Law.
In this view, 'factors' (such as capital movements, sector movements) and 'interval arbitrage' models are calculated and presented based on publicly available, objective data. What the software does is merely simplify the information processing process without making subjective 'value judgments' or 'investment decisions'. Users conduct transactions based on their own settings, and the software only executes automatically when conditions are triggered, serving as a tool for user-determined decisions, rather than the company making decisions on behalf of the user.
If merely using data analysis and automation technology constitutes a crime, then all AI software or financial data terminals providing similar information screening and quantitative tools (such as certain features of Tonghuashun and Dongfang Caifu) on the market could also be deemed illegal operations, which is clearly unreasonable.
So why is the business engaged in by S Company identified as illegal business operations?
3. Why the court determined it constituted illegal business operations: Disassembly of S Company's business model
From the perspective of S Company's business model, the actual situation far exceeds the simple sale of AI stock trading software. The company not only provides strategy services such as 'interval arbitrage' but also divides its stock trading robot software into different tiers of membership services, specifically 8800 yuan and 28800 yuan VIP members. Through these membership services, S Company charges clients an 'interface usage service fee', which is a fee for 'illegally accessing brokers' trading channels', ultimately accumulating profits of over 3 million yuan.
Specifically, members of the 8800 yuan tier will receive services such as 'interval arbitrage'. 'Interval arbitrage' refers to the software analyzing the rise and fall of stocks over several days, calculating reference values, recommending individual stock buying and selling timing to clients, or directly providing trading strategies to assist clients in quantitative investment. If clients are unsatisfied with the software’s preset algorithms and trading strategies, they can obtain further analysis and recommendations through customer service personnel.
Members of the 28800 yuan tier can set their own investment 'track' and fund allocation plan based on the data, models, and parameters provided by the company. Once preset conditions are triggered, the software will automatically assist clients in executing buy/sell operations.
Therefore, the reason S Company constitutes illegal business operations is:
The service priced at 8800 yuan essentially provides clients with specific investment advice—recommending individual stock buying and selling timing and providing strategic trading. Customer service recommends based on empirical analysis, and this service directly points to 'telling clients what to buy and when to buy', with the core being providing investment advice, which has completely exceeded the scope of information intermediary.
The service priced at 28800 yuan directly replaces clients in executing trades by combining the data and models provided by the company. The software's automated execution function essentially plays the role of trading channel and instruction execution, which is the core of brokerage business.
The combination of the two allows S Company to bypass the role of licensed brokers and directly complete the core securities trading and investment consulting services that should be conducted by licensed institutions such as securities companies and investment consulting firms.
4. The compliance boundaries of AI stock trading software: distinguishing between crime and non-crime
So based on the above case, does it mean that related individuals or companies, without obtaining (business license for securities and futures), (qualification for securities investment fund sales), and other qualifications, selling AI stock trading software is itself a behavior prohibited by law? Certainly not.
In practice, the core of the dispute is not whether AI is used or whether quantification is done, but rather: what is ultimately outputted by the products and services—is it data and tools, or 'investment advice and trading execution'? In other words, has it already crossed the boundary from neutral technical services to securities investment consulting or securities business operations?
Combining S Company's business model, a comparison can be made on the following three levels:
1. Functional output aspect:
Among S Company's services, the 'interval arbitrage' model and customer service recommendations output clear instructions on 'when to buy and sell which stock'. This completes the transition from 'what is' (data facts) to 'how to do' (investment decisions), which is the essence of investment consulting business.
According to (Interim Provisions on Strengthening the Regulation of Securities Investment Consulting Business Using 'Stock Recommendation Software' (2020 Second Revision)) Article 1: Software products, software tools, or terminal devices that have the function of summarizing securities information or historical data statistics of securities investment varieties but do not possess functions such as 'providing specific investment analysis opinions on securities investment varieties, forecasting price trends, suggesting variety choices, and actual buy/sell recommendations' do not belong to 'stock recommendation software'.
Therefore, if an AI tool only provides securities information summary, historical data statistics, capital flow, or sector movement displays without providing specific investment analysis opinions, trend forecasts, variety selection suggestions, and buy/sell recommendations, it is closer to a neutral information tool.
2. Business logic aspect:
S Company's business model involves charging high and tiered membership service fees, with the fee basis being 'providing profitable investment advice and trading convenience'. From the user's perspective, they are paying for 'who can quickly complete profit-making operations'—this is essentially selling investment consulting services and trading channel services.
In contrast, if a product only sells software once or charges a data subscription fee, with its main value being the information tool itself and not using 'profitability' as the core selling point, it is usually closer to the business logic of technical services or information services.
3. Operational closed-loop aspect:
Among S Company's services, especially in the 28800 yuan tier, after users set the conditions, the software fully automates the entire process from decision trigger to order execution. This has partially replaced the roles of investors and brokerage agents.
In contrast, if a product only provides analytical tools and users still need to judge and place orders themselves, with the decision-making and execution processes not replaced by the platform, then its overall risk boundary is relatively controllable.
5. The support of 'black technology' leads to multiple penalties
S Company's software can quickly help clients conduct automated stock trading because it purchased the external program from 'hacker' Han and integrated it into its company program, deploying it on the company server, allowing S Company's account to bypass the relevant technical protection measures of the Tongdaxin software (market terminal provider), call its trading channel, illegally access the brokerage server, and conduct automated stock trading, thus charging clients a service fee for interface usage.
In this regard, Han was convicted of providing programs that intrude into computer information systems and sentenced to three years. The actual controller of S Company, Zhong, and the technical head, Kong, were both subject to multiple penalties for illegal business operations and copyright infringement.
The reason is that under the compliance framework, the basic structure of securities trading should be:
(I still prefer the most original drawing method)
Exchanges are responsible for trading rules and matching mechanisms, while brokers, as the only legal operating entities, establish client relationships with investors, who participate in securities trading through brokers. Service providers like Tongdaxin, which provide market and trading terminal support, do not engage in securities business operations nor establish securities service relationships with investors.
Investors first establish a trading relationship with S Company, purchasing so-called 'interval arbitrage', 'stock trading robots', and other products and services. S Company then illegally accesses the brokerage trading channel by cracking the Tongdaxin interface program and bypassing normal technical protection measures, completing automated trading operations, and continuously charging fees to investors for profit.
In this structure, brokers and Tongdaxin are not the organizers or beneficiaries of the business but rather become 'utilized subjects' in its technical path. The entity that truly builds the business model, controls the trading path, forms a charging closed loop, and profits from it is always S Company itself.
6. Risk alert for AI entrepreneurs and programmers:
This case of a technology company being sentenced for 'AI stock trading software' essentially exposes the blurred boundaries of compliance for many AI entrepreneurial projects, quantitative trading products, and financial technology tools.
For AI entrepreneurs, company leaders, technical partners, programmers, and even product managers, if you are engaged in or plan to enter related fields, it is advisable to prudently evaluate your own products and business models to ensure they do not touch the regulatory boundaries of securities business, avoiding being misjudged as investment consulting or disguised securities business operations.