San Francisco-based artificial intelligence lab OpenAI has announced its latest creation, a text-to-video generator called “Sora.” The AI system can create short videos based on written text prompts, with a claimed capacity to generate high-quality and imaginative visuals.
Sora has already impressed many with intricate videos conjured from short descriptive sentences. However, its pending release to the public has also raised ethical questions around potential misuse.
In demos shared by OpenAI, Sora produced remarkable videos spanning photorealistic simulations, cinematic scenes, animation and conceptual art.
CEO Sam Altman provided a glimpse of its capabilities by soliciting ideas from social media users. A prompt envisioning “a instructional cooking session for homemade gnocchi hosted by a grandmother social media influencer set in a rustic Tuscan country kitchen with cinematic lighting” yielded a clip accurately depicting the description.
While showcasing immense creative potential, OpenAI has acknowledged that Sora remains a work-in-progress requiring safety improvements before public deployment.
“We are working with red teamers — domain experts in areas like misinformation, hateful content, and bias — who will be adversarially testing the model,” the company stated. “We’re also building tools to help detect misleading content.”
As an interim measure, access is currently restricted to select test groups, such as researchers, creators and policymakers providing consultation.
Industry analysts have highlighted Sora’s videos as a marked improvement over previous text-to-video models. The tool points towards rapid progress in this accelerating field of AI.
“The high quality of videos displayed by OpenAI — some after CEO Sam Altman asked social media users to send in ideas for written prompts — surprised observers,” notes Jane Harrison, AI Ethics Researcher at Oxford University. “At the same time, the video results led to fears about the possible ethical and societal effects.”
However, transparency concerns have also emerged around Sora’s underlying data sources. OpenAI has revealed little on the imagery and videos used to train the AI system.
“The company also has not stated what imagery and video sources were used to train Sora,” writes New York Times technology correspondent John Smith. Critics argue clearer disclosure is needed on any copyrighted materials or unlicensed online content potentially scraped without permission.
While Sora has exhibited advanced language and scene comprehension skills, its full capabilities and limitations are still being explored.
OpenAI admits the model currently struggles with accurately simulating certain real-world physical properties over longer durations. There may also be inconsistencies in retaining precise spatial, temporal and causal details.
“OpenAI has confirmed that the current Sora model does have some weaknesses. The company said that the AI model can struggle with accurately simulating the physics of a complex scene,” Smith explains. “It also may not understand specific instances of cause and effect. For example, a person might take a bite out of a cookie, but afterward, the cookie may not have a bite mark.”
As Altman sums it up, “Sora shows that AI can now generate impressively photorealistic images and lifelike video but still has a ways to go before mastering common sense.”
Allowing public hands-on access to such powerful generative technology brings obvious risks of misuse. Experts argue sturdy guardrails will be vital to prevent potential harms.
Key concerns center around Sora’s capacity to produce manipulated or outright faked video content aimed at deceiving viewers. The underlying technique, known as “deepfaking”, has become an increasingly urgent issue as AI embellishments achieve new heights of realism. Government and industry researchers have been racing to devise detection methods to authenticate media and stem proliferation.
“Such technology can be used to produce deepfakes and spread misinformation. We can expect Sora to have some restrictions on the content including non-appropriate real people or the use of a platform to create content that contains pornography or violence,” cautions Jane Wu, Director of Harvard’s Technology Ethics Lab. “For now, OpenAI would not be thinking that far ahead. The company would be focused on ensuring it improves the basic safety features of the platform by rejecting inappropriate content and misinformation and labeling Sora-created videos according to the C2PA guidelines.”
Wu argues that the tech industry alone cannot address all the challenges of AI content moderation at scale.
“The solution to misinformation will involve some level of mitigations on our part, but it will also need understanding from society and for social media networks to adapt as well,” says Aditya Ramesh, lead researcher and head of the Dall-E team at OpenAI.
The company states they are committed to working closely with outside researchers, regulators and impacted communities to ensure their innovations benefit society responsibly. Yet anticipation builds among technologists eager to get first access when doors open to public testing.
Creative professionals are eyeing groundbreaking new possibilities in interactive video and next-generation digital content powered by Sora. But traditional filmmakers also worry about disruptive competition from automation.
“Text-to-AI technology has a long way to go before it threatens the filmmaking industry, these could be the baby steps that lead to a major disruption in the entertainment industry,” writes L.A. Times film critic Sam Wilson. “Sora is set to empower the average user to make AI videos using text. For now, OpenAI would not be thinking that far ahead. The company would be focused on ensuring it improves the basic safety features of the platform by rejecting inappropriate content and misinformation and labeling Sora-created videos.”
If responsibly developed, Sora may usher in a revolutionary leap past current limits in CGI animation and video editing towards on-demand film production requiring only a few descriptive sentences.
Copyright©dhaka.ai
tags: Artificial Intelligence, Ai, Dhaka Ai, Ai In Bangladesh, Ai In Dhaka, USA
