Use of alternative financial data via AI
Alternative financial data of a company is any data about the company not pro- duced by that company. Examples of such data that can benefit from improved collection and aggregation using AI models include stock discussions on social media, product reviews, and satellite imagery. The use of alternative financial data, enabled by the deployment of AI models, may introduce biases and generalization issues due to shorter shelf-life and vary- ing quality (e.g., shorter time series, smaller sample sizes, and dubious claims) due to its origins from various sources, posing financial tail risks (i.e., tail-end of a probability distribution), where the price of a company changes dramatically [4].
ENTITY
1 - Human
INTENT
2 - Unintentional
TIMING
2 - Post-deployment
Risk ID
mit1189
Domain lineage
5. Human-Computer Interaction
5.1 > Overreliance and unsafe use
Mitigation strategy
1. Establish Robust Data Governance and Quality Control Implement a comprehensive framework for vetting, structuring, and cleaning alternative data sources prior to their use in AI models. This must include data preprocessing techniques like rebalancing and introducing fairness constraints during model optimization to systematically mitigate inherent biases, addressing the risk of generalization issues and varying quality from non-traditional sources. 2. Mandate Continuous Monitoring and Model Audits Establish protocols for real-time monitoring of AI model performance and continuous bias detection post-deployment. This involves mandatory, independent out-of-sample backtesting of model predictions and cross-checking alternative data insights against reliable traditional financial datasets to verify findings and prevent the amplification of errors or dubious claims. 3. Apply Advanced Techniques for Tail Risk Management Adopt sophisticated Machine Learning models, such as Tail-GANs or ensemble approaches, to enhance the accuracy of extreme market scenario simulation and tail risk estimation. Furthermore, employ structural analysis and "inverting assumptions" techniques in scenario design to anticipate cascading systemic risks that statistical extrapolation alone would fail to capture, thus addressing the potential for dramatic price changes.