What is a possible ethical resolution to address fair use of copyrighted material in AI training (select all good answers)?
A. Only using data that is explicitly labeled as free to use.
B. Negotiating agreements with content owners for data usage.
C. Implementing a system to track and credit original content creators.
D. Limiting AI training to non-commercial data sources.
Correct answer: A, B, C.
Explanation: With (A.), this avoids any copyright issues while having a large enough test set to not overfit it. With (B.), this expands the database for the AI, creating an expansive, overfit-proof set for the AI to train on, at the expense of increased costs. (C.) is a good option because it expands the database while increasing transparency, but it may be complex to implement. (D.), however, may not be a great option because commercial data sources make a large portion oof available sets, risking overfitting or not enough training. Therefore, answer (A.), (B.), and (C.) are correct.