Toward the Humane Use of AI
As part of Integrity Week, Floris Mreijen (Director of DSI) is writing a guest column for the International Chamber of Commerce (ICC) this year. From October 26 through 30, the 10th edition of Integrity Week will take place, with this year’s highly topical theme: “Leading with Integrity in a Digital Age.”
As a partner of Integrity Week, DSI draws attention to the impact of technological innovation on our field. Artificial intelligence is rapidly transforming the investment sector, but how do we ensure that ethics, transparency, and human judgment remain central? In the article below, Floris shares his insights on how humans and algorithms work together and why integrity remains the decisive factor, especially in the digital age.
Curious about the full program or the resources for this anniversary edition? Check out the 2026 Toolbox on the Week of Integrity website.
AI requires investment professionals to have different skills
DSI maps out in the CTRL SHFT dialogue paper how artificial intelligence is driving fundamental changes in the investment industry. According to CBS figures from 2024, 37.4% of companies in financial services were already using AI technologies at that time, a sharp increase compared to the year before. By now, this percentage is likely significantly higher.
This development confronts professionals with a difficult dilemma. Who is responsible when an algorithm co-determines investment advice? And how does someone remain professionally competent when traditional knowledge is no longer sufficient?
Responsibility remains with the professional, even with AI
The principle of personal responsibility requires professionals to always stand behind the quality of their advice. The CTRL SHFT document emphasizes, however, that AI often functions as a so-called black box: outputs are available, but the underlying logic remains opaque.
In practical terms, this means that investment professionals cannot rely blindly on algorithmic output. They must actively take control, ask critical questions, and assess outcomes against the client’s interest. For example, an AI tool recommends a specific portfolio allocation. The professional verifies whether this choice aligns with the individual client’s risk tolerance and objectives, rather than adopting the recommendation without question.
Professional competence now also means being able to question AI
Professional competence is taking on a new meaning. Where professionals previously mainly required substantive knowledge, they must now also understand how AI models operate and where potential risks or biases may arise. The adviser becomes a director as well.
The analysis in the document shows that this is not an optional skill, but a core competency. Professionals who do not make this shift risk falling behind. In practice, this means:
- Developing insight into how the algorithms used actually work;
- Learning to recognize when data bias may distort outcomes;
- Keeping up with technological developments through continuous education.
The client’s interest remains paramount, including when technology increases efficiency. AI can accelerate processes but must not lead to advice that loses the human dimension or becomes opaque to the client. Transparency requires professionals to explain where and why AI has been used in the advisory process.
“Technology is evolving rapidly, but making judgments remains a human task.”
Time for reflection and action
The CTRL SHFT document is not an endpoint, but a starting point for discussion. It helps organizations and professionals to jointly reflect on the implications of AI for integrity and professional competence. This reflection is urgent: technology is developing rapidly, but judgment remains human work.
Professionals ask themselves to what extent they currently rely on automated systems without verifying the underlying logic. And whether they are willing to step back when AI moves faster than their own judgment.
Organizations that take these questions seriously can download the document via the DSI website or contact DSI for more information. Those who invest in knowledge and dialogue now will be better prepared for the changes AI brings.
DSI and AI
At DSI, we are actively engaging with the topic of AI. The dialogue paper serves as input for discussions with the sector, but the question of what the adviser of the future looks like also directly affects our organization. Will we still be certifying advisers in the future? Or algorithms? Or will our organization become obsolete soon? We are not only investing in facilitating the dialogue but have also started applying AI ourselves.
We developed an AI-powered content platform built on an agent-based architecture that combines a centralized knowledge base, configurable AI agents, and structured workflows to generate high quality content at scale[4]. The system helps us write articles based on our own disciplinary case law, unlocking our internal archive. In addition, we can read online news articles with our system and use them as a basis for new content for our certified professionals to learn from.
A news report on investment fraud can thus be transformed into an article from which professionals in the industry can learn. Naturally, this is done with respect for copyright; one of our agents is programmed as a “guard agent” and legal adviser.
This article is also partly AI-generated.
As an exercise, I asked our bot to write a piece using content from our website, with the dialogue paper on AI as its foundation. From the subheading “DSI and AI” onward, I wrote the text myself, after checking the first part for content and consistency and adding just a few lines. This brings me to my main point: we are moving toward a world in which AI makes our lives easier in many areas. The distinction between human and machine is becoming less clear. The role of humans, however, is not. Transparency about where an algorithm is or has been used is crucial, as illustrated in the first part of this article. Integrity is, of course, also a central theme.
As humans, we must remain alert to AI output. In my conversations with professionals in the investment industry, this is a consistent signal I receive. Where AI is being experimented with, less experienced or less well-trained professionals may assume that a sound recommendation has been generated. More experienced advisers, however, still regularly identify flaws in these automated recommendations. At the same time, it is a fact that AI will take over many tasks. Everyone must prepare for this and consider what the organization, the professional, and the client relationship will look like in the future.
Human knowledge remains important for the time being. In addition, professionals must become more aware of their soft skills. Tools are helpful for calculations, but assessing a client’s personal situation and, for example, family dynamics in the case of long-term invested assets, remains a distinctly human responsibility. Perhaps that is a good thing.
