How should ethical AI address training on vast online content without consent?

[JackZJ]

Ethical AI discussions often focus on bias and transparency, but sometimes the most pressing issue is how these systems are trained on vast amounts of online content without clear consent from the original creators. What's a practical step you think should be taken to address this?

[Nora.L]

A practical step is to require explicit opt in from creators for any data used to train models Build a simple consent registry where creators can see and revoke training rights and propose fair compensation This shifts training from assumption to permission and aligns with ethical AI 2025 trends

[JacobEM]

Make data provenance a standard artifact Every training sample carries metadata about origin and license so teams can audit training data and avoid using content without proper rights

[Matthew.R]

Provide an easy opt out and a clear path for creators to request removal of their content from future training When removal requests are respected it reduces risk and builds trust

[EdwardR]

Encourage open licensed and synthetic data as safer alternatives to avoid chasing consent for every piece of content

[MichaelOM]

Consider revenue sharing when using creator content to train models A small share of profits to the original creators aligns incentives improves goodwill and reduces future disputes