There is a suggestion to have companies commit to training their models on licensed data to ensure artist protections. Initiatives like the Content Authenticity Initiative embed metadata in content to track creation details and allow for tagging content with 'do not train' credentials. Laws and regulations alone are considered insufficient to protect artists, and active efforts are needed to safeguard creator rights. The speaker reflects on the use of quizzes for password authentication that unknowingly contributed to training AI, highlighting the need for transparency in data usage practices.