CFA Institute, the global association of investment professionals, has released new research and an accompanying ethical decision framework to motivate the evolution of ethical practices in the development of artificial intelligence (AI) technologies in investment management.
The new research, Ethics and Artificial Intelligence in Investment Management, a Framework for Professionals, recognizes
the spectrum of issues brought about by AI tools and big data in
investing, and sets out questions for professionals and investment teams
to consider when working with AI technologies at each step of the AI
workflow. The paper combines fundamental ethical principles with the
applicability of relevant professional standards as set out in the CFA Institute Code of Ethics and Standards of Professional Conduct to offer a decision framework to guide the development of responsible AI applications in investment management.
The
research also identifies organization-level requisites for AI to be
successfully used in a variety of applications that include investment
analysis, portfolio management, risk management, trading, automated
advice, and client onboarding.
Rhodri Preece, CFA, Senior Head, Research, CFA Institute comments:
“AI
adoption offers significant potential benefits, yet also entails several
risks. The expansion of data sources and availability of AI tools to
harness big data can improve investment decision making but can also
introduce more complexity in investing. With ever more data sources and
more complex decision-making algorithms, the application of AI must
prompt firms and professionals to re-examine the span of ethical
considerations in investing. The framework aims to support professionals
in the ethical design, development, and deployment of AI tools.”
Key Takeaways:
- Investment organizations must establish a culture
conducive to client-centric AI innovation, and incorporate a robust risk
management and governance framework that includes regular model
testing. A talent development programme to ensure the acquisition of
appropriate knowledge, skills, and abilities within investment teams can
give firms an edge.
- Ethical considerations regarding the use of AI in
investment management include the integrity of data, the accuracy and
validity of models, transparency and interpretability of algorithms, and
accountability structures.
- AI models should avoid bias and excessive complexity or
opacity so that they can be interpreted and understood by all relevant
stakeholders. They should yield fair and accurate outcomes.
Interpretability methods play an important role supporting model
transparency.
- Regular model testing and review should be part of the
governance framework surrounding the use of AI to ensure that
applications perform and evolve as expected.
- Model development and evaluation should consider the
existence or emergence of biases in data or in the way that models learn
from features, the interpretability of the contribution of features to
the outcome, the fairness and accuracy of outcomes, and the ongoing
suitability of models to client needs.
- An ethical decision framework sets out the relevant
questions professionals and investment teams should consider when
working with AI technologies at each step of the AI workflow.
Rhodri Preece, CFA, continues:
“Instilling an ethical decision framework in AI-driven
investment processes is critical to ensure applications serve the best
interests of clients. Given the complexity of AI projects, senior
leadership must establish a strategic vision and ethical culture for AI
development within the organization. While the use of AI in investment
management is still relatively formative, it is appropriate that we
examine the ethical aspects of AI implementation to guide future
developments responsibly. We offer our research and framework as
guidance for investment professionals to advance the industry’s efforts
to incorporate AI responsibly.”
CFA
© CFA Institute
Key
Hover over the blue highlighted
text to view the acronym meaning
Hover
over these icons for more information
Comments:
No Comments for this Article