OpenAI wasn’t expecting Sora’s copyright drama

Following the rapid rollout of its video generation application, Sora, OpenAI has pivoted its copyright strategy by transitioning from an opt-out model to one that gives intellectual property holders direct control over their content.

This decision followed widespread public backlash when users quickly generated unauthorized content featuring trademarked characters and created controversial deepfakes. CEO Sam Altman admitted the company did not anticipate the intense public reaction or the complex demand for nuanced permissions regarding personal likenesses. In response, OpenAI has introduced new features allowing individuals to restrict how their digital avatars, or "cameos," are utilized, such as banning their likeness from political or offensive contexts.

Beyond user-control challenges, the platform faces ongoing safety concerns as internet users have already discovered ways to bypass watermarks designed to identify AI-generated media. Despite these vulnerabilities and criticism of OpenAI's gradual rollout approach, Altman defended the strategy as a necessary step for society to adapt to highly realistic synthetic media. Furthermore, the immense popularity of Sora has accelerated OpenAI's focus on securing massive computational power. To address these resource constraints, the firm is aggressively pursuing infrastructure initiatives, including major data center projects and a new partnership with chipmaker AMD, to sustain its highly demanding technology stack.