Record summary
A quick snapshot of what this page covers.
Risk profile
How this risk is described and categorized.
"This is an emerging category, with more cases prone to appear as the use of generative AI tools–such as Stable Diffusion, Midjourney, or ChatGPT–becomes more widespread. Some content creators are already suing for the appropriation of their work to train AI algorithms without a request for permission or compensation. Perhaps even more damaging cases will appear as developers increasingly ask chatbots or assistants like CoPilot for ready-to-use computer code. Even if these AI tools have learned only from open-source software (OSS) projects, which is not a given, there are still serious issues to consider, as not all OSS licenses are equal, and some are incompatible with others, meaning that it is illegal to mix them in the same product. Even worse, some licenses, such as GPL, are viral, meaning that any code that uses a GPL component must legally be made available under that same license. In the past, companies have suffered injunctions or been forced to make their proprietary source code available because of carelessly using a GPL library."
Suggested mitigations
Defenses that may help with related attacks.
Source
Research source for this risk, when available.
Included resource
Navigating the Landscape of AI Ethics and Responsibility
Original source
MIT AI Risk Repository
Open the public repository used for AI risk records and taxonomy fields.
