OpenAI announced the launch of Sora 2, a new artificial intelligence video generator, on Tuesday. The platform now includes a social feed enabling users to share realistic videos. However, shortly after the release, numerous videos began appearing in the feed and on other social media platforms, featuring copyrighted characters in inappropriate contexts, as well as graphic imagery depicting violence and racism. OpenAI’s policies prohibit content that promotes violence or causes harm, raising questions about enforcement.
Prompts and video clips reviewed by various sources included unsettling scenarios such as bomb threats and mass shootings set against familiar urban landscapes. One particularly alarming example depicted children in war-torn areas discussing their homes being destroyed by conflict. Other clips emulated news reports featuring fictional responses to civil unrest, while some videos unambiguously referenced racist slogans.
Despite being invite-only and not broadly available, Sora 2 quickly escalated to the top of Apple’s App Store shortly after its limited release. The head of Sora, Bill Peebles, remarked on the flood of creative content generated by users. However, experts contend that the realistic nature of such AI-generated videos poses challenges for truth verification, potentially facilitating misinformation, fraud, and harassment.
OpenAI’s CEO, Sam Altman, acknowledged both excitement and concern regarding the potential misuse of the technology. He noted that while the company has implemented measures to safeguard against harmful content, instances of copyrighted characters and misleading narratives have already emerged.
Before launching Sora 2, OpenAI approached studios and talent agencies to inform them that they could opt-out if they wanted to protect their intellectual property. The company indicated it would work with rights holders to address copyright issues but has yet to implement comprehensive preventive measures for all users.
Questions remain regarding the effectiveness of these safeguards as misuse and copyright infringements continue to proliferate on the platform.
Source: https://www.theguardian.com/us-news/2025/oct/04/openai-sora-violence-racism