ChatGPT's new feature can help with stock picking! How to regulate AI when it enters the investment field?

ChatGPT’s new feature can help with stock picking! How to regulate AI when it enters the investment field?

China Times (www.chinatimes.net.cn) reporter Qi Meng Zhang Zhi reporting from Shenzhen

OpenAI recently announced on Twitter that it will provide networking capabilities and multiple plug-ins to all ChatGPT Plus users this week. It is worth noting that there is a plug-in called Portfolio Pilot, which can compare the expected returns of two stocks. It will also score the investment portfolio uploaded by the user and give improvement suggestions. It can also query the stock price, price-earnings ratio, and company price of listed companies. News etc.

Last week, the "AI digital clone" of Gu Jia, chief executive of China Merchants Securities Media and co-leader of the TMT group, also officially appeared in front of the public. Expert interviews show that although the application of AI technology in the investment field has certain potential and impact, there are also challenges and problems.

An Guangyong, an expert from the Credit Management Committee of the All-China Mergers and Acquisitions Association, said in an interview with a reporter from the China Times that AI technology has doubts about the quality and accuracy of data for investment suggestions and decision-making, and lacks the ability to explain. At the same time, AI is in the analysis The impact of irrational behavior cannot be clearly captured when it comes to markets and trends, and it may also trigger a series of ethical and legal issues. Jiang Han, a senior researcher at Pangu Think Tank, also pointed out to this reporter that the issue of investment risk responsibility after AI decision-making still needs to be resolved.

Regarding relevant regulations, lawyer Fu Jian, director of Henan Zejin Law Firm, pointed out to this reporter, “There are already some relevant regulations for the supervision of financial institutions’ use of artificial intelligence technology, such as the “New Asset Management Regulations”. However, AI brings challenges to financial supervision and other aspects, and legislation still needs to be further improved."

In this regard, lawyer Zhao Tianjiao of Shanghai Landi Law Firm suggested that the regulation of artificial intelligence can be supervised from the perspective of full cycle, hierarchical classification, technical certification, and organizational structure. “It can be done in data source collection, annotation and model Full-process supervision is carried out in terms of training, deployment of applications, result output, and reprocessing; the main responsibilities can be divided into source data and infrastructure providers, model algorithm and other technology developers, application service providers, etc."

AI enters the financial investment field

Among the many OpenAI plug-ins, the most eye-catching is Portfolio Pilot. With ChatGPT after this plug-in, you can compare the expected returns of two stocks, score the investment portfolio uploaded by the user and give improvement suggestions, and query the stock price, price-earnings ratio, company news, etc. of listed companies.

According to information, the plug-in was released as early as early May. It is claimed to be the first officially certified "investment plug-in" and was co-created by AI software entrepreneur Alex Harmsen. He has said the software is backed by hedge fund models.

Gu Jia serves as the chief of China Merchants Securities Media and co-leader of the TMT group, and his "AI digital clone" has recently attracted widespread attention on the Internet, which also shows that artificial intelligence is gradually entering the investment field. It is explained that this AI digital avatar not only has the appearance and voice of Gu Jia, but can also interact and generate in multiple modes. These abilities are derived from a large model. She can stay up 24 hours a day, transform into hundreds of millions of clones, and appear at road shows, press conferences, research report interpretations, analyst conference calls, and wherever clients need her. According to the news, starting from this week, Gu Jia will use his digital representative status to share the weekly media and Internet industry research report.

In fact, in recent years, digital transformation has become an important direction for innovation in all walks of life. In the financial field, not only CICC Wealth Management has launched "digital employees", but Industrial Securities and Tongtai Fund have also launched digital clones of virtual employees and "digital clones" of fund managers.

In October last year, Jinn, the "digital employee" of CICC Wealth Management, made his debut as a digital consultant assistant to accompany users to deepen their understanding of investment knowledge; in December last year, the "digital clone" of Wang Xiaogen, the fund manager of Tongtai Fund Company, also Official debut. Its main job responsibility is also to accompany and serve investors.

In February this year, Industrial Securities launched a fully-realized digital avatar virtual employee "Xiaozhi" based on its own employees and integrating "broadcasting services". It is positioned to integrate market information broadcasting, investor education and business The digital intelligence service officer that integrates consulting and management services can not only create one-click content production links and produce videos in seconds to cope with fluctuations in securities prices, but can also support team cloud query, management, sharing, editing and co-creation.

Opportunities and challenges coexist

In this regard, the interviewed experts said that the current application of AI technology in the investment field has certain potential and impact.

In An Guangyong’s view, AI technology can generate intelligent investment suggestions through faster data processing and decision-making capabilities, and can help investors better manage risks.

Jiang Han also pointed out that applying AI technology to the investment field can not only save a lot of human resources, but also help investors make up for deficiencies and omissions in financial knowledge. He said: “Investment requires a lot of basic work such as data analysis. Currently, these tasks are completed by junior analysts or interns. If AI is used, a lot of human resources can be saved; secondly, many investors have a low level of financial knowledge. Relatively limited, AI can play a role in filling in gaps or popularizing science.”

However, it should be noted that there are also some challenges and problems in the application of AI technology in the investment field.

In An Guangyong’s view, AI technology has doubts about the data quality and accuracy of investment recommendations and decision-making, and lacks the ability to explain. At the same time, AI cannot clearly capture the impact of irrational behavior when analyzing markets and trends. In addition, a series of ethical and legal issues may arise.

He said that in terms of data, AI investment recommendations and decisions rely on a large amount of data input. If the data quality is poor or inaccurate, it will lead to wrong predictions and investment decisions; in terms of decision-making, because AI is usually based on black Decisions are provided in the form of a box, and it is difficult to explain the logic and decision-making basis behind them. This makes it difficult for investors to understand and trust AI recommendations. At the same time, the investment market will be affected by emotions, irrational behavior, and group behavior. AI is analyzing The market and when predicting trends may not be able to accurately capture these irrational factors, leading to erroneous predictions and decisions.

"In terms of ethics, the application of AI in the investment field has triggered a series of ethical and legal issues, such as privacy protection, data security, responsibility for artificial intelligence decision-making, etc.; in terms of risk management, AI should not only focus on decision-making Benefits while ignoring risks may lead to portfolio instability and potential losses." An Guangyong believes.

Jiang Han also pointed out that who should bear the investment risks after AI decision-making is also an issue that needs to be resolved.

Regarding the above issues, An Guangyong suggested that high-quality data sources should be used for data cleaning and verification, and accurate models should be established; transparent AI models and algorithms can be developed to provide explanations and interpretability, so that investments can Investors can understand the basis for decision-making; can incorporate behavioral finance principles and psychological factors into the model to better understand investor behavior; can develop risk management models, use historical data and market indicators to evaluate risks, and take appropriate risk management Control strategy; establish an appropriate regulatory framework to ensure that the use of AI complies with ethical and legal requirements and protects the rights and interests of users.

Regulations need to be improved

Specifically in terms of regulations, regarding the compliance issues of "AI digital human avatars" in the financial industry, lawyer Zhao Tianjiao said that it needs to be analyzed based on the specific conditions of this type of product. From the perspective of behavioral attributes, if the AI ​​digital clone does not directly provide investment advisory services to customers and make profits, for example, it only provides general investment information compilation, then from a theoretical analysis, it only needs to meet my country's requirements for the dissemination of Internet information content and artificial intelligence technology. regulatory requirements. my country's regulations and standards on artificial intelligence supervision are still being improved. The "In-depth Synthesis Management Regulations on Internet Information Services" have come into effect, and the "Measures for the Management of Generative Artificial Intelligence Services (Draft for Comments)" are still being solicited for opinions.

"On the other hand, if the behavior engaged in by the AI ​​digital clone already falls within the scope of 'investment advisory' business as stipulated by law, in addition to the above requirements, the relevant provisions of my country's "Interim Regulations on Securities Investment Advisory Business" should also be complied with. For example, relevant institutions should accept the supervision and management of the China Securities Regulatory Commission and the Securities Association, and relevant personnel should be registered as securities investment consultants with the Securities Association of China." Zhao Tianjiao said.

"At the same time, from the perspective of the nature of the subject, the subject itself that uses AI digital clones to provide services should be financial institutions such as securities companies and securities investment consulting institutions that meet regulatory requirements. At the same time, institutions that provide technical services should fulfill the financial technology requirements For the relevant supervision and management obligations of products and services, such as product certification, external evaluation, self-regulatory management and other requirements, please refer to the central bank’s "FinTech (FinTech) Development Plan (2019-2021)" and other documents." Zhao Tianjiao further pointed out.

In terms of relevant regulations, Fu Jian pointed out that there are already some relevant regulations. “The New Asset Management Regulations require that financial institutions using artificial intelligence technology to carry out asset management business shall strictly abide by relevant investor suitability, investment scope, Information disclosure, risk isolation and other general provisions prohibit the use of artificial intelligence business to exaggerate the promotion of asset management products or mislead investors. AI has brought challenges to financial supervision and other fields, so relevant legislation needs to be further improved."

So, when there are disputes or problems with the information and opinions expressed by the AI ​​digital avatar of a financial institution, how to hold them accountable?

Fu Jian said that the current legal regulations on relevant responsibilities are not perfect, but AI digital clones are equivalent to a special investment advisory service business carried out by qualified practitioners and institutions. Therefore, if disputes arise, we will provide Institutions and practitioners serving AI clones should be the responsible entities.

AI brings huge challenges to financial supervision. It is reported that my country has not yet established an industry-specific and hierarchical supervision system for the application of artificial intelligence.

In this regard, Zhao Tianjiao suggested that the regulation of artificial intelligence can be supervised from the perspective of full cycle, hierarchical classification, technical certification, and organizational structure.

He suggested that the whole process of data source collection, annotation and model training, deployment and application, result output, re-processing, etc. should be supervised; the main responsibilities should be stratified, which can be divided into source data and infrastructure providers, Model algorithm and other technology developers, application service providers, etc.; classify application scenarios into information consulting, auxiliary decision-making, etc.

"Evaluate and standardize technologies, conduct long-term security assessments of new technologies and new applications, and update relevant policies on this basis, improve relevant data standards, technical standards, service standards, penalty standards, and develop products and services Certification, provide safety service products; in terms of organizational structure, establish regulatory agencies and corresponding management measures from central to local levels, encourage industry organizations, professional institutions, etc. to participate in collaborative governance, establish relevant committees within market entities, and implement self-discipline management; the judicial department improves relevant Recognition standards." Zhao Tianjiao said.

The above is the detailed content of ChatGPT's new feature can help with stock picking! How to regulate AI when it enters the investment field?. For more information, please follow other related articles on the PHP Chinese website!