The following is a guest article by Susan Mathews, Principal & Founder of Madigail Consulting, LLC a securities compliance consultant providing research and ethics compliance, training and external GC and CCO services to young Wall Street firms. Susan was previously an Enforcement Attorney at the SEC and was the Regulator Liaison for the independent research portion of the SEC’s Global Research Analyst Settlement.
AI is now a widely available, inexpensive tool for drafting basic legal documents. While I have spent over 30 years drafting ethics and compliance frameworks for Wall Street participants, I recently discovered that properly worded inquiries asked of ChatGPT or other legal-focused AI sites such as AILawyer can result in fairly comprehensive internal policies and procedures. Have I been replaced by technology?
I spent several days asking AI sites to tell me what compliance and ethics issues I need to be concerned about and to draft internal policies for several different types of Wall Street entities, including broker-dealers, expert networks, and equity research firms. Then I submitted detailed inquiries to the two sites I found most user-friendly and responsive: ChatGPT and AILawyer (collectively “AI”). You may be surprised at what I found.
For purposes of this article, I focus on internal ethics and compliance issues for an equity research firm. This firm is typically an unregulated Wall Street entity with a small number of employees; however, since its clients usually demand “best practices,” research firms must devote time and resources to build a comprehensive ethics and compliance framework to attract and retain clients. I put myself in the role of an internal manager who wants to draft a basic Code of Conduct to save outside counsel costs.
First Submitted Question: What are the most important compliance and ethics issues I need to focus on?
First, I asked AI what the most important ethics and compliance issues are that an equity research firm should be worried about. ChatGPT was only able to list five issues: insider trading, conflicts of interest, research quality, compliance with regulations, and fair dealing. Given it is not a legal-focused site, this is not a surprising result. However, the same query to AILawyer resulted in some truly bizarre results. It was able to list some expected topics: objectivity and independence, confidentiality, insider trading, conflicts, and record-keeping. We can agree that those are important issues facing equity research today.
However, AILawyer also included unhelpful and ambiguous concepts such as: compliance with laws and regulations, personal conduct, professionalism, and “disclosure requirements related to conflict of interest, compensation and other relevant information.” Regulators cringe at over-generalized provisions with no specific direction for employees to follow:
- What is the difference between personal conduct and professionalism? Without giving specific behavioral direction, such as “be a good listener,” “don’t interrupt” or “don’t use foul language,” these overly general terms cannot be put into practice.
- Which laws and regulations do employees need to follow?
- What are the conflict-of-interest requirements that relate to compensation? I’m unaware of any unless the AI is referring to outside employment?
Regulators demand code language provide clear instruction about human behavior.
Second Submitted Question: Draft an Internal Ethics and Compliance Code that Governs Employee Conduct.
ChatGPT prefaced its code with a general concept: research independence and objectivity. It states: “Our research is based solely on factual analysis, and we do not allow any conflicts of interest to influence our research.” (See full text at end of article) This is an impressive opening, as it demonstrates that ChatGPT knows what regulators have been focusing on for the past several years and understands the importance of conflict-free research. I was also happy to see that the ChatGPT code also includes personal trading restrictions which forbid employees from trading securities in companies the firm covers or provides research on. Sadly, the remaining code failed to impress.
ChatGPT initially generated a very basic code with a few necessary concepts listed, but nothing defined or explained. It mentions confidentiality, fair dealing, and conflicts disclosure, but does not articulate their meaning or how to apply these concepts to employee behavior. AILawyer was also able to list basic code concepts, but also lacked explanatory language. Unlike ChatGPT, AILawyer included essential insider trading and supervision. Consequently, it appears a combo of both sites could provide a skeletal ethics and compliance code. But don’t toss your attorney yet!
Neither site generated a code that will survive regulator scrutiny or the demands of your clients, regardless of what type of Wall Street entity you are. Vital concepts were missing from both AI codes, including whistleblowing, social media policy, information controls, data protection and privacy. Even more troubling – – when a concept is included, it is not adequately explained or included in an example applying the concept to employee behavior. A code cannot just state that “conflicts of interest” should be disclosed to clients without explaining what a conflict is; the types of conflicts the firm should be screening for; and how these conflicts might affect the objectivity of their research.
Neither site specified the actual standards the firm is required to meet. For example, both codes contain a general statement on maintaining records of business-related activities, but do not mention how many years records should be maintained, or how a “business-related activity” is defined. Regulators expect these policies to state the number of years records must be preserved as required by the applicable law, regulation, or rule (three years), and legal concepts to be defined.
Moreover, regulators require an entity’s policies to be “closely tailored” to their particular business. Consequently, if I include a conflict-of-interest provision in a code for a research provider that covers the energy sector, I might include an example in which a family member owns stock in Exxon, as this would closely mirror a realistic potential conflict situation within that firm.
In sum, both AI tools failed to put me out of a job. When I drilled down on each general concept by asking a minimum of 4 follow-up questions, each site was able to provide more specific guidance and language that could be added to the code. However, if you are not an experienced compliance attorney, you would not know what follow-up questions to ask. The three salient take-aways from this exercise are:
- If you use AI to draft your code, understand the lack of differentiation: Neither AI tool distinguished among different types of equity research firms that exist on the Street. Consequently, they were drafted in such general terms that rendered them virtually useless. A human attorney knows to ask differentiating questions before drafting:
- what type(s) of research does the firm do,
- does the firm also conduct other business activities such as investment banking,
- what are the sources of the firm’s research?
Basic inquiries like these allow an attorney to target where potential violative or conflicted actions might occur.
- Laughable Definitions: Both tools defined a term using the same term. They failed to provide definitions or guidance without repeating the same word in their attempt to define the word. For example:
- “Professional Conduct:” employees must conduct themselves professionally at all times.
- “Confidentiality:” employees must maintain the confidentiality of client information.
- Outdated information and just quitting: when I asked for recent guidance, cases, or regulations, ChatGPT responded: “My training data only goes up to September 2021.”
AILawyer wasn’t even polite enough to give a response – it just stopped responding altogether.
In closing, I am happy to report AI has not rendered attorneys obsolete! In fact, when I queried these AI sites with difficult legal questions that do not have clear answers, both responded with ridiculously conservative legal approaches and instructions to consult with your legal and compliance counsel. I will explore AI’s approaches to insider trading in my next article.
For more news about alternative data, investment research, and investment related FinTech, click here to sign up for Integrity’s free ResearchWatch Weekly newsletter, or click here to subscribe to Integrity’s proprietary ResearchWatch news and commentary service.
EXHIBIT
Independent Equity Research Provider Code of Ethics and Compliance Generated by ChatGPT
As an independent equity research provider, we are committed to upholding the highest standards of ethics and compliance. Our mission is to provide accurate and unbiased research to help our clients make informed investment decisions. In order to achieve this goal, we have established the following code of ethics and compliance:
1. Independence and Objectivity: We are committed to providing unbiased and independent research to our clients. Our research is based solely on factual analysis, and we do not allow any conflicts of interest to influence our research.
2. Confidentiality: We respect the confidentiality of all information provided by our clients and do not disclose any confidential information to any third parties without prior written consent.
3. Compliance with Laws and Regulations: We comply with all applicable laws and regulations, including those related to securities regulations and anti-bribery laws. We maintain appropriate licenses and registrations required by the regulatory bodies.
4. Fair Dealing: We treat all clients and investors fairly and impartially, providing all clients with access to the same research reports and recommendations.
5. Disclosure of Conflicts of Interest: We disclose any conflicts of interest that may exist in our research and ensure that our clients are informed of such conflicts. We do not accept any compensation, gifts, or other benefits that could influence our research.
6. Accuracy and Transparency: We strive to provide accurate and transparent research to our clients. We clearly disclose our research methodologies and sources of information, and we do not make any exaggerated or misleading claims about the companies we cover.
7. Personal Trading Restrictions: Our employees are subject to personal trading restrictions to avoid any potential conflicts of interest. We do not permit employees to trade securities in companies that we cover or provide research on.
8. Record Keeping: We maintain accurate and complete records of our research, client interactions, and any other activities related to our business.
9. Training and Education: We provide ongoing training and education to our employees to ensure they are knowledgeable about the latest regulations, ethical practices, and industry developments.
10. Reporting Violations: We encourage our employees to report any suspected violations of this code of ethics and compliance. We investigate all reports of violations and take appropriate disciplinary action when necessary.
By adhering to this code of ethics and compliance, we maintain our integrity, earn the trust of our clients, and contribute to the integrity of the financial markets.