tag

OpenAI has announced that it's changing how ChatGPT's Voice mode works on the web and app.The company stated that as part of the update, users can interact with ChatGPT Voice directly within the ongoing chat, allowing users to see the transcript of conversations with OpenAI's AI model, as well as visual elements that illustrate what ChatGPT is talking about.It was added that, with the new update, users can now start a voice conversation by tapping the "sound waves” icon next to the text field. Unlike the previous interface, which relied on a separate design filled with moving circles, Voice mode is now integrated directly into the conversation itself, making it easier to switch between speaking and typing.This integration between visual elements and voice responses comes as a natural extension of ChatGPT’s multimodal capabilities. Users can already issue voice requests supported by images or video clips, and the model’s voice responses provide the same level of visual integration.

Chinese tech giant Alibaba reported a quarterly revenue bump on Tuesday, after shares rose on the blockbuster launch of a new artificial intelligence app downloaded more than 10mn times in a week.Analysts said the early success of the company's redesigned Qwen app positions it as a potential competitor to Chinese rival DeepSeek and OpenAI's ChatGPT.Alibaba said on Tuesday revenue in the three months ending September 30 totalled 247.8bn yuan ($35bn).That was a 5% jump on the 236.5bn yuan logged in the same quarter last year and beat analyst estimates of 245.2bn yuan."We have entered into an investment phase to build long-term strategic value in AI technologies and infrastructure," said Eddie Wu, Alibaba's chief executive officer."Robust AI demand further accelerated our Cloud Intelligence Group business, with revenue up 34 % and AI-related product revenue achieving triple-digit year-over-year growth for the ninth consecutive quarter," Wu said.Net income attributable to ordinary shareholders for the quarter fell to 21bn yuan, down 52% on-year, the company said.Alibaba runs some of China's biggest online shopping platforms and wants to become a major player in the global race to develop AI technology.The Hangzhou-based company said in February it would spend at least 380bn yuan ($53bn) on artificial intelligence and cloud computing over the next three years.Wu said in September the company plans to increase that spending further.This approach has seen Alibaba's share price on Wall Street surge around 90% from this time a year ago — despite fears that AI-related stocks are overvalued worldwide and could crash.Its US shares closed 5% higher on Monday after the company said the Qwen app had surpassed 10mn downloads within the first week of its beta launch.Crystal Li and Tommy Wong of China Merchants Securities said last week that Qwen's launch was "supported by Alibaba's prolonged investment and cutting-edge capabilities in foundational models" for artificial intelligence.It could pave the way for the adoption of AI agents — programmes that use chatbots to do the work humans do online, such as buying a plane ticket or adding events to a calendar."Compared to ChatGPT's direct access to third-party apps, we believe Alibaba's self-developed comprehensive product ecosystem and free-to-use policy provide competitive advantages for Qwen app," Li and Wong wrote.Emily Jarvie of Proactive Investors also noted ahead of Tuesday's earnings release that "China is a key market for Qwen, as OpenAI's ChatGPT is not available"."Its rapid adoption makes it one of the fastest-growing AI apps in China," Jarvie said.Alibaba has also been in the headlines recently for other reasons, namely geopolitical tensions between China and the US.

OpenAI officially announced the launch of its AI-powered video creation app, Sora, for Android. The app includes a "Cameos" feature that allows users to create videos of themselves performing various activities using their real-life features and artificial intelligence (AI). These videos can be shared and interacted with by other users, a strategic move by the company to strengthen its position in the short-form video market.OpenAI said that it plans to add new features to the app, including the creation of animated videos, alongside basic video editing tools such as merging multiple clips to produce longer, multi-scene videos. It also plans to develop tools that allow users to customize suggested content to show content from specific accounts, giving them greater control over what they watch.The Sora app, which was released for iOS devices last September, is now available on the Google Play Store for Android devices.

OpenAI has introduced a new feature that enables users to access and use third-party applications directly within ChatGPT, in a move designed to transform the popular AI assistant into an integrated productivity platform capable of carrying out complex tasks in collaboration with online services. According to OpenAI, users can now open compatible applications during conversations, allowing ChatGPT to provide context, reasoning, and guidance while the external app executes specific actions. At launch, the available applications include Booking.com, Canva, Coursera, Expedia, Figma, Spotify, and Zillow, while additional apps such as DoorDash, OpenTable, Target, and Uber are expected to follow in the coming weeks. The company said the rollout builds on OpenAI's recent efforts to make ChatGPT more interactive, following the introduction of in-chat purchasing features earlier this year. In its official statement, OpenAI announced the release of a Software Development Kit (SDK), which allows developers to begin building experimental applications today. Later this year, developers will be able to submit their apps for review and publication through a dedicated directory that will enable users to browse, install, and interact with available tools. OpenAI explained that the new capabilities are based on the Model Context Protocol (MCP), an open standard that connects ChatGPT with external services, facilitating seamless collaboration between users and third-party applications. Demonstrations shared by OpenAI showed how users could, for example, design a poster through Canva and then ask ChatGPT to generate a presentation using Figma, all within the same chat interface. The company also highlighted use cases such as searching for flights with Expedia or viewing home listings via Zillow, illustrating how everyday tasks can now be completed without leaving the chat. OpenAI emphasized that privacy and data control remain central to the experience. When a user first connects to an app, ChatGPT will display clear prompts explaining what information is being shared and for what purpose. The company stated that the feature will gradually roll out to users outside the European Union across the Free, Go, Plus, and Pro plans, with EU availability to follow later this year. OpenAI said the update reflects its broader vision of evolving ChatGPT from a text-based assistant into a "chatable" platform, one where users, developers, and services collaborate seamlessly in a unified workspace.

OpenAI is preparing to launch a standalone social app that resembles the TikTok experience, coinciding with the launch of its video generation AI model, Sora 2. According to a report published by Wired magazine, the app will feature a vertical video feed and swipe-to-scroll navigation. However, it will be limited to AI-generated content only, with no option to upload clips from a user's phone, and videos within the app are limited to 10 seconds or less, while the model's limit outside the platform remains unclear, Wired said. The report added that the app will integrate an identity verification feature that allows users to confirm their likeness created by the model, enabling others to tag users and use their likeness when they go to remix one of their videos, with the user receiving a notification immediately whenever their likeness is used by someone else, even if it is not posted. Regarding intellectual property owners, the app will not include copyrighted material, the report said. In this context, the Wall Street Journal said that OpenAI will require property owners to explicitly ask OpenAI not to include their copyright material in Sora 2 content. By incorporating a social aspect into the Sora model, OpenAI may seek to create its own community, making users less likely to migrate to competing platforms. Experts expect the app to be an early test of OpenAI's strategy of combining generative models with everyday user experiences, potentially opening the door to new forms of interaction between AI and social content. OpenAI recently launched ChatGPT Pulse for Pro subscribers, a tool that generates quick daily updates in the form of visual cards based on the user's personal data, including chat history and connected apps like Google Calendar and Gmail. These cards include personalized suggestions such as exercise plans, meal ideas, language lessons, and more.