AI in Capital Markets: Top 5 Ethical Risks

Published January 6, 2023 | 2 min read

How can we ensure artificial intelligence is bringing real value to clients?

AI in Capital Markets: Top 5 Ethical Risks
As artifcial intelligence (AI) becomes more embedded in fnancial
decision-making, how can we ensure it’s bringing real value to clients?
These 5 key concerns must be addressed:
- AI is often a black box – opaque and complex 64% of IT professionals
to explain and understand. say that being able to
explain how their AI arrives
- But explainability is critical for trust, end-users at different decisions is
must feel confdent engaging with predictions. important to their business.1
2 Bias
- Data, and people, can be unintentionally biased and
that can adversely infuence AI models.
Between 3.4% and - There are documented cases where racial bias has
38.6% of data used been detected in AI systems3
, for example.
to inform certain AI
systems is biased.2 - Discriminatory bias can signifcantly erode trust.
Competitive market dynamics
- Automation and using AI to generate deeper
trading insights is opening up new competitive
dynamics in markets.
- Greater competition could lead to issues
stemming from collusion, market dominance,
or mergers.5
Since 2017 at least 60 countries
worldwide have enacted AIrelated guidelines.4
4 Robustness
- Robustness, or reliability, is essential for AI used
in capital markets.
- Robustness refers to the AI model’s sensitivity
Only 50% of people trust companies that use to any inputs that might result in incorrect
AI as much as they trust other companies.6 predictions or results.
5 Data security and privacy
- AI is heavily reliant on large quantities of data, and
without proper controls, data can be corrupted.
Almost 20% of IT experts
- Addressing privacy concerns while leveraging believe data security is
large datasets is also a challenge – risks include a major concern when it
the exposure of personal identifable information. comes to AI systems.7
Building the future
RBC Capital Markets is committed to building better AI.
Discover more about our approach to responsible and
explainable algorithms in our article: Responsible
and Explainable AI: Exploring the Future of Trading.

Download PDF