The almost reliable way to prove the work is yours is showing a timestamp or some document that you owned the content pre-AI boom. Now it's original creativity with spin-offs on the AI era. One starts an idea and there would be more versions of that idea where some of them go viral or better than the original.
The process isn't new because we've been seeing that with old inventions. It's just that the process just became faster that we notice it more often.
The future is likely involuntarily accepting that whatever you share online will be put into training models used by the platform or someone else.
Authors and publishers (including Andrea Bartz and Charles Graeber) accused Anthropic of using over 7 million pirated books downloaded from shadow libraries like Library Genesis to train its Claude AI models.
The legal ruling here was it was fair use if the content used for training was acquired legally. So they got penalized for using content downloaded from pirated sources.
Most wouldn't even notice their accounts have a setting has a default turned on for the platform to use their content for AI training as seen on Instagram/Facebook/X. It's not like these platforms highlight this subtle fact or most people becoming meticulous about reading the terms and conditions.
So if your content gets posted under these platforms, it may automatically assume your consent was given and it's legal.
Small fish don't have the financial powers to pursue legal actions against multi-billion dollar corporations. The compensation for violating you copyright may not be worth it. It's easier or cost-effective to pay the fines. And if there more entities out there that are privately training their AI so trying to sue them all isn't feasible. Once the content is out there, it's not like they can "untrain" their AI from your work readily.
Thanks for your time.