Record summary
A quick snapshot of what this page covers.
Techniques0Attack methods connected to this risk.
Mitigations0Defenses that may help with related attacks.
Domain6. Socioeconomic and EnvironmentalThe broad risk area this belongs to.
Risk profile
How this risk is described and categorized.
Domain6. Socioeconomic and Environmental
Subdomain6.6 > Environmental harm
Entity2 - AI
Intent2 - Unintentional
Timing3 - Other
CategoryAutomation, Access and Environmental Harms
SubcategoryEnvironmental harms from operation LMs
Suggested mitigations
Defenses that may help with related attacks.
No propagated mitigations. No defense is available through the connected attack methods.
Source
Research source for this risk, when available.
Included resource
Ethical and social risks of harm from language models
Original source
MIT AI Risk Repository
Open the public repository used for AI risk records and taxonomy fields.