Question19 The article "Regulating Artificial Intelligence in Finance: Putting the Human in the Loop" argues that external governance (traditional financial supervision) is insufficient for AI risks. Instead, it advocates for strengthening internal governance. What is the core reason provided for this preference? Internal governance is less costly to implement than external supervision. External supervisors lack the legal authority to regulate AI in finance. Financial institutions prefer internal governance to avoid public scrutiny. The severe information asymmetry and data dependency make external supervision challenging. ResetMaximum marks: 1 Flag question undefinedSingle choice
A
Internal governance is less costly to implement than external supervision.
B
External supervisors lack the legal authority to regulate AI in finance.
C
Financial institutions prefer internal governance to avoid public scrutiny.
D
The severe information asymmetry and data dependency make external supervision challenging.

Question

Question19 The article "Regulating Artificial Intelligence in Finance: Putting the Human in the Loop" argues that external governance (traditional financial supervision) is insufficient for AI risks. Instead, it advocates for strengthening internal governance. What is the core reason provided for this preference? Internal governance is less costly to implement than external supervision. External supervisors lack the legal authority to regulate AI in finance. Financial institutions prefer internal governance to avoid public scrutiny. The severe information asymmetry and data dependency make external supervision challenging. ResetMaximum marks: 1 Flag question undefinedSingle choice

A

Internal governance is less costly to implement than external supervision.

B

External supervisors lack the legal authority to regulate AI in finance.

C

Financial institutions prefer internal governance to avoid public scrutiny.

D

The severe information asymmetry and data dependency make external supervision challenging.

BlackTom AI · Accepted Answer

The severe information asymmetry and data dependency make external supervision challenging.

1. Relevant knowledge points and concepts
- Information asymmetry: firms possess far more, and often proprietary, knowledge about their AI models, training data, and operational details than external regulators.
- Data dependency: AI systems’ behavior depends heavily on training datasets, data pipelines, and ongoing data flows that are internal and dynamic.
- Internal governance: firm-level controls (model validation, documentation, human-in-the-loop, monitoring) that are closest to the data and systems.
- External supervision: traditional regulatory oversight that is limited by legal access, technical resources, confidentiality concerns, and observational gaps.

2. Step-by-step reasoning
Step 1: AI model behavior is determined by proprietary training data and complex pipelines that change frequently. Step 2: External supervisors cannot continuously observe model inputs, retraining events, or operational configurations without deep, ongoing access. Step 3: This creates severe information asymmetry: regulators see outputs (or limited reports) but lack full visibility into causes. Step 4: Because effective risk management requires access to data lineage, model internals, and validation artifacts, the entity with direct control (the financial institution) must be responsible for strong internal governance. Step 5: Thus strengthening internal governance is the practical remedy to AI risk, while external supervision remains necessary but insufficient.

3. Intermediate “steps” / illustrations
- Regulators rarely have real-time access to training datasets or feature engineering details.
- Proprietary models are often “black boxes,” and performance can drift as data distribution changes, requiring internal monitoring and rapid remediation capabilities.

4. Why the correct answer is correct
The article emphasizes that the combination of hidden, proprietary data and complex, evolving AI systems produces information gaps that make it impractical for external supervisors alone to detect and manage AI risks. Therefore internal governance — where the data and models reside — is the crucial lever.

5. Why other options are incorrect
- “Less costly”: Cost is not the core argument; feasibility due to information access is.
- “Lack legal authority”: The article focuses on practical observability and data dependency, not primarily on legal authority.
- “Avoid public scrutiny”: That suggests motive, but the article’s reason is structural (asymmetry/data dependency), not institutional preference.

6. Conclusion
For these reasons, the severe information asymmetry and data dependency make external supervision challenging — confirming the provided correct answer.