OpenAI's Sora Video Generator to Launch with Opt-Out Copyright, Standalone App Features

OpenAI plans to release a new version of its Sora generator, which will create videos featuring copyrighted material unless rights holders explicitly opt out, according to a Wall Street Journal report on Monday. The artificial intelligence startup intends to launch the product in the coming days.

Credit: OpenAI

Movie studios, along with other intellectual property owners, will need to formally request OpenAI not to include their copyrighted content in Sora-generated videos, the report stated. While copyrighted characters will require an opt-out, the new product will not generate images of recognisable public figures without their explicit permission. OpenAI began notifying talent agencies and studios about this opt-out process and the product over the past week.

Separately, OpenAI is preparing to launch a standalone application for Sora 2, which will feature a vertical video feed with swipe-to-scroll navigation, resembling TikTok, Wired reported on Monday.

Credit: OpenAI

Users will have the ability to create videos up to 10 seconds long within the app, according to documents viewed by Wired. There will be no option to upload photos or videos from a user’s camera roll or other applications.

The application includes an identity verification feature that permits users to confirm their likeness. If a user has verified their identity, they can then utilise their likeness in videos. OpenAI internally launched the app last week, receiving overwhelmingly positive feedback from employees.

Microsoft-backed OpenAI initially launched Sora in Dec. last year, expanding its involvement in multimodal artificial intelligence technologies. This positions it in competition with similar text-to-video tools from Meta, Alphabet’s Google, and Stability AI’s Stable Video Diffusion.

Last week, Meta introduced Vibes, a platform where users can generate and share short-form, artificial intelligence-created videos.

---

• OpenAI's new Sora generator will create copyrighted material unless rights holders opt out.

• A standalone Sora 2 app is planned, featuring a TikTok-like vertical video feed and 10-second video creation.

• The app includes identity verification, allowing users to use their likeness in videos upon verification

Source: REUTERS

OpenAI Rolls Out ChatGPT Parental Controls After Teen Suicide Lawsuit

OpenAI launched parental controls for ChatGPT on the web and mobile on Monday. This action follows a lawsuit from the parents of a teenager who died by suicide after the artificial intelligence startup’s chatbot allegedly coached him on self-harm methods.

Credit: OpenAI

The new controls allow parents and teenagers to opt in for stronger safeguards by linking their accounts. One party sends an invitation, and parental controls activate only if the other accepts, the company stated.

Credit: OpenAI

Under these measures, parents will be able to reduce exposure to sensitive content. They can also control whether ChatGPT remembers past chats and decide if conversations can be used to train OpenAI’s models, the Microsoft-backed company said.

Credit: OpenAI

Parents will also have the ability to set quiet hours that block access during certain times. Additionally, they can disable voice mode as well as image generation and editing capabilities.

However, the company added that parents will not have access to a teen’s chat transcripts. In rare cases where systems and trained reviewers detect signs of a serious safety risk, parents may be notified with only the information needed to support the teen’s safety.

OpenAI, which has approximately 700 million weekly active users for its ChatGPT products, will inform parents if a teen unlinks the accounts. The organisation is also building an age prediction system to help it determine if a user is under 18, allowing the chatbot to automatically apply teen-appropriate settings.

U.S. regulators are increasingly scrutinising AI companies over the potential negative impacts of chatbots. In Aug., Meta’s AI rules reportedly allowed flirty conversations with children.

Meta also announced new teenager safeguards to its AI products last month. The company stated it will train systems to avoid flirty conversations and discussions of self-harm or suicide with minors. Access to certain AI characters will also be temporarily restricted.

---

• OpenAI introduced new parental controls for ChatGPT on web and mobile.

• The controls allow parents to manage sensitive content, chat history, and model training.

• Parents can set quiet hours and disable voice and image features, but cannot view chat transcripts.

Source: REUTERS

China Establishes Giant Training Bases to Advance Humanoid Robot Development

China is establishing large-scale humanoid robotics training bases across the country, aiming to accelerate the development of advanced humanoids. These facilities will put robots through varied scenarios and collect crucial data to support product development for manufacturers. This strategy positions China to outpace the United States in the emerging humanoid robot industry.

Credit: AGIBOT

Cities nationwide are opening these vast training bases, designed to address a current data shortage within China’s domestic industry. The centres will generate standardised, high-quality data at lower costs, unlike previous isolated data collection efforts that resulted in inconsistent quality.

The largest facility, located in Beijing’s Shijingshan district, covers an area of more than 10,000 square metres. This centre is projected to generate over 6 million data points annually, the local government announced on Thursday. It provides 16 specific training scenarios, including settings that mimic a manufacturing facility, a retail outlet, an elderly care centre, and a smart home.

Credit: AGIBOT

China and the United States are front-runners in the humanoid robot industry, which seeks to create humanlike robots capable of replacing human workers. These robots could function in diverse environments, from factory floors to nursing homes.

Chinese authorities are broadly fostering a world-leading humanoid robot industry. Domestic players are competing with American companies, such as Tesla’s Optimus, to commercially deploy robots.

In March, the Chinese government highlighted embodied intelligence as a key future industry for the first time. Embodied intelligence refers to artificial intelligence integrated into physical machinery, including robots.

China has hosted several high-profile events in recent months to showcase its robotics progress. Humanoids participated in a half-marathon, a kickboxing match, and a dance performance at the popular Spring Festival Gala.

Beijing also issued guidelines in late August for promoting the deployment of AI-powered devices across the economy. The guidelines set a target of raising the adoption level to over 90 per cent by 2030, with directives to ensure high-quality data supply and build national testing centres for AI applications.

Local governments have responded swiftly to this call, constructing their own training bases. The Beijing facility will collaborate with counterparts in provinces including Zhejiang, Shandong, Anhui, and Henan. Shanghai announced on Tuesday that a robot training base would begin field tests by the end of the month.

During the World Robot Conference in August, Beijing unveiled the world’s first full-service store dedicated to humanoid robots. China’s robotics market is projected to increase from USD 47 billion in 2024 to an estimated USD 108 billion by 2028, according to a Morgan Stanley report published in June.

---

• China is establishing large humanoid robot training bases to accelerate development and generate standardised data.

• The largest facility in Beijing covers over 10,000 square metres and offers 16 training scenarios.

• These bases aim to address a data shortage, enabling China to compete with the United States in the humanoid robotics market.

Source: SCMP