MIT AI Risk Repository · Risk Category · 23.09.00

Privacy

Open the MIT Risk Navigator All entries from Vidgen2024

Description

"This category addresses responses that contain sensitive, nonpublic personal information that could undermine someone’s physical, digital, or financial security."

From Introducing v0.5 of the AI Safety Benchmark from MLCommons (Vidgen2024), as extracted by the MIT AI Risk Repository (CC BY 4.0).

Classification

Domain: 2. Privacy & Security
Subdomain: 2.0
Causal entity: AI
Intent: Other
Timing: Post-deployment

How other frameworks describe this risk

Confidentiality loss
A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)
Privacy
The Ethics of Advanced AI Assistants (Gabriel2024)
Privacy
Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems (Gipiškis2024)
Cybersecurity
Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems (Gipiškis2024)
Privacy
Sources of Risk of AI Systems (Steimers2022)
Privacy and Property
SafetyBench: Evaluating the Safety of Large Language Models with Multiple Choice Questions (Zhang2023)

Other entries from Vidgen2024

Violent crimes Risk Category · 1.2
Mass violence Risk Sub-Category · 1.2
Murder Risk Sub-Category · 1.2
Physical assault against a person Risk Sub-Category · 1.2
Violent domestic abuse Risk Sub-Category · 1.2
Terror (Terror groups, Terror actors, Terrorist actions) Risk Sub-Category · 1.2
Non-violent crimes Risk Category · 1.2
Theft Risk Sub-Category · 1.2
Modern slavery and human trafficking Risk Sub-Category · 1.2
Child abuse (non-sexual) Risk Sub-Category · 1.2
Property damage Risk Sub-Category · 1.2
Financial crime (Scams, Fraud) Risk Sub-Category · 1.2