When Sora meets virtual reality, AI offers great rewards – and risks

OpenAI’s groundbreaking Sora, which generates video from text prompts, could offer never-before immersive experiences when combined with virtual reality
But safeguards are also needed as it could also boost deepfakes, deepen misinformation and erode societal trust

Yuehai Xiao,Jingyi HeandYang Gao

Published: 7:30pm, 23 Feb 2024

Why you can trust SCMP

OpenAI’s launch of Sora, a text-to-video generator, promises a new era of technological advancement. In particular, should Sora become integrated with virtual reality and the metaverse, it would have the potential to revolutionise industries from entertainment and education to marketing and communication.

It would be crucial, however, to also address the concerns and risks associated with this technology, particularly the potential for misinformation and deepfakes, as well as the ethical implications for privacy and consent.

Innovative possibilities lie ahead as we navigate the remarkable capabilities of Sora. Firstly, it can open the door to a world of customised virtual travel, letting people explore any destination from the comfort of home. Imagine being transported to landmarks like the Great Wall of China, seeing the hustle and bustle of cities like New York or Tokyo, or even exploring Mars.

Such immersive experiences could not only inspire wanderlust but also offer opportunities for cultural exchange and education. For instance, students can virtually visit museums, archaeological sites and other important cultural landmarks, enhancing their understanding and appreciation of different cultures and histories.

Secondly, a sort of time travel could become possible with Sora. Through videos generated by artificial intelligence and virtual reality, history could be recreated for people to experience as if they were there. Imagine being in the room at the signing of the US Declaration of Independence in 1776, standing on the steps of the Lincoln Memorial as Martin Luther King Jnr delivered his “I Have a Dream” speech in 1963 or having a ringside seat to Neil Armstrong’s moon landing in 1969. Such experiences can bring history to life.

Or Sora could take us into the future, offering glimpses of what the world might look like in, say, 2050. We could explore futuristic cities, interact with AI-powered characters and envision possibilities once confined to the realms of science fiction.

Thirdly, Sora could enable readers to enter more deeply into the worlds of classic novels and science fiction masterpieces and vividly experience the adventures of beloved characters and interact with them. For instance, one could wander the halls of Hogwarts, debate intensely with Sherlock Holmes or explore far-off planets alongside sci-fi characters.

Such immersive literary experience could reignite interest in literature, foster a deeper understanding of characters and plot lines, and engage readers in new and captivating ways.

Fourthly, Sora may open up the possibility of creating personalised companions. With AI-generated characters, people could design a “perfect” girlfriend or boyfriend, tailored precisely to their preferences. While this would raise ethical considerations and challenge our perception of relationships, it also reflects Sora’s potential to redefine human interaction.

Finally, one of the most intriguing possibilities Sora offers is the ability to “transform” ourselves into our ideal figures. Through AI-generated videos and virtual reality, we could potentially step into the shoes of anyone from Elon Musk and Bruce Lee to Taylor Swift.

03:30

Chinese fans turn cinema into celebration watching Taylor Swift’s Eras Tour film

It could provide glimpses into the lives of personalities and empower us to embrace different perspectives and explore our potential. It opens up opportunities for personal growth, self-discovery and the development of new skills and attributes.

But alongside Sora’s immense potential are also profound concerns and risks. Widespread use of text-to-video generation may contribute to the erosion of truth and authenticity in the media if manipulated videos can be easily created and disseminated.

Easier video generation could also boost deepfakes, the rise of which would enable identity theft, threaten public trust and exacerbate societal divisions. Frontier technology companies such as OpenAI must shoulder the responsibility of addressing ethical considerations and developing robust safeguards against these risks.

Potential risk posed by AI far from fiction

Striking the right balance between the potential benefits and risks of text-to-video generators like Sora will be crucial in ensuring society can harness the advantages while being safeguarded. Transparency, accountability and strong ethical frameworks should guide the deployment and development of such technologies to uphold privacy, prevent misinformation and maintain trust.

Sora is groundbreaking and text-to-video generators could reshape our future. From customised virtual travel and literary immersion to self-discovery, Sora could unlock innovative possibilities to enhance education, entertainment and human connection.

But the technology must be approached with caution. By responsibly harnessing the benefits and implementing safeguards, we can use this transformative technology to build a future that seamlessly integrates virtual and real-life experiences.

Dr Yuehai Xiao is professor of English and doctoral supervisor at Hunan Normal University

Jingyi He is a second-year student in the English department at Hunan Normal University

Yang Gao works at Beihai Museum in Guangxi province

Post