A company's AI risk team is asked to assess the risk of deploying a generative AI system for creating marketing copy. The legal team flags a concern about the system potentially reproducing copyrighted material from its training data. What is the MOST appropriate technical control to mitigate this risk?
Select an answer to reveal the explanation.
Short Explanation
You need a technical control that catches copied text before it goes live, so output filters and similarity checks against known copyrighted works are the right move. They automatically reduce the risk of publishing infringing marketing copy.
Full Explanation
The most appropriate technical control is implementing output filters and similarity-checking tools that compare generated content against a database of known copyrighted works before publication, because it provides direct, automated protection against copyright infringement and addresses the specific legal risk identified. Instructing users not to submit copyrighted material as prompts is a behavioral control that can be bypassed and doesn't address the model's tendency to reproduce training data. Requiring all AI-generated marketing copy to be reviewed and approved by a human editor is a good complementary control but relies on reviewer knowledge of all copyrighted works. Adding a disclaimer to all AI-generated content stating that it may contain copyrighted material does not prevent infringement or reduce legal liability.