Get paid to talk and train AI

How do the $150 hourly rate for recording and the distribution bonus work together? Is it settled based on effective hours or by the number of recordings? It's best to clarify the billing basis upfront so we don't end up arguing about it later.

For referrals, it would be best to clarify upfront whose contract the freelancer would be under, which party pays the compensation, and who owns the recording copyright; otherwise, no matter how high the hourly rate is, it's hard to feel confident participating when the boundaries of your rights aren't clear.

What scenarios will the recordings be used for, and can I request that unused material be deleted?

The real execution challenge lies in screening and quality control: simply "finding people to talk" makes it hard to guarantee voice quality and consistency, and downstream labeling, denoising, and accent coverage all carry hidden costs.

I actually think recruitments like this should focus more on how the voice data will be used and how long it will be retained. A high hourly rate easily draws people in, but a speaker's voiceprint is sensitive information—without a clear scope of use and an opt-out mechanism, participants can hardly assess the risk. Looked at another way, transparent rules matter more than the pay figure.