MIT AI Risk Repository · Risk Sub-Category · 62.16.14
Benchmark Inaccuracy (Benchmarks may not accurately evaluate capabilities)
Category: Model Evaluations
Description
"Benchmarks of AI systems can both underestimate and overestimate the capa- bilities of those AI systems. Underestimates can happen if an evaluation is not comprehensive enough, if the benchmark is saturated by existing models, or if the capabilities in question depend on a complicated setup, such as realistic computer programming tasks. Overestimates of capabilities can occur if an AI system is trained or fine-tuned on the contents of the benchmark, leading to overfitting."
From Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems (Gipiškis2024), as extracted by the MIT AI Risk Repository (CC BY 4.0).
Classification
- Subdomain
- 6.5 Governance failure
- Causal entity
- Human
- Intent
- Unintentional
- Timing
- Pre-deployment
Subdomain definition: Inadequate regulatory frameworks and oversight mechanisms failing to keep pace with AI development, leading to ineffective governance and the inability to manage AI risks appropriately.
Real-world incidents in this subdomain
- Nippon Life Alleged ChatGPT Practiced Law Without a License in Illinois Disability Case
- OpenAI Allegedly Did Not Alert RCMP After ChatGPT Flagged Violent Chats Before British Columbia School Shooting
- Remotely Operated Taser-Armed Drones Proposed by Taser Manufacturer as Defense for School Shootings in the US
How other frameworks describe this risk
- Liability issues in case of accidents
- Mobility
- Faster scientific progress makes it harder for governance to keep pace with development
- Type 1: Diffusion of responsibility
- Products Liability Law
- Difficult to develop metrics for evaluating benefits or harms caused by AI assistants
- Institutional responsibilities
- Governance - Regulation