Latest news with #SynthID
Indian Express
2 days ago
- Indian Express
Google Gemini can now illustrate storybooks based on your prompts and images
Google Gemini just rolled out a new feature that lets you bring your ideas to life. It personalises and illustrates storybooks based on your ideas with read-aloud narration. With just a description of any narrative that the user can think of, Gemini will create a customised 10-page book complete with original artwork and audio. You can ask Gemini to source ideas from your images and documents for a genuinely unique touch. Whether it's pixel art, comics, claymation, crochet, or even colouring books, you may realise your idea in over 45 languages. If users are having trouble helping a child understand a complex topic, with the assistance of Gemini, they could create a story that explains the complex topic. For example, if you want to explain the solar system to a five-year-old, Gemini would create a story explaining that. Or if you are having trouble teaching young children about being kind, the chatbot will teach them through storytelling with customisable characters. The feature can also bring personal artwork to life and recite stories. If you upload your kids' drawings and prompt the Gemini to write a creative storybook according to the drawing, Google claims that it can execute it. The feature could also turn your photos into stories. In recent times, Gemini has also integrated the feature which lets you turn your photos into videos using their video generative model, Veo 3. Users can transform their memories into eight-second AI-generated videos. According to Google, the new tool lets users express their creativity by enabling them to do things like add motion to natural scenes, bring sketches and paintings to life, and animate common objects. In addition to an invisible SynthID digital watermark, Google Gemini will automatically add a visible watermark to your photographs when you convert them to movies to show that they were created with artificial intelligence. Additionally, you can use the thumbs-up or thumbs-down button to provide Gemini comments on these Veo 3-generated videos. The new photo-to-video feature, which is now only available to Google AI Pro and Ultra members, will be made available to users in a few countries starting on July 11. If it isn't available on your account at the moment, you may have to wait.
AFP
2 days ago
- Science
- AFP
Social media users fall for AI-generated clip of flooding in northern India
"Rain in Uttar Pradesh submerged many houses," reads the Instagram post published on July 4, 2025. It also shares a clip that appears to show a group of people on the roof of a building and buffaloes walking in floodwater, with Hindi-language text on the clip largely repeating the false claim. Image Screenshot of the false post taken on July 25, 2025, with the red X mark added by AFP The posts surfaced as torrential rain hit Himachal Pradesh and Uttarakhand -- two states that neighbour Uttar Pradesh -- in June and July, killing at least 69 people (archived link). Floods later affected Uttar Pradesh in August, displacing more than 11,000 people and damaging over 300 homes (archived link). India's annual monsoon season from June to September offers respite from the intense summer heat and is crucial for replenishing water supplies, but also brings widespread death and destruction. The false clip also spread on Facebook and X, and some users appear to believe the clip was filmed in Uttar Pradesh. One user wrote, "No one can escape the fierce form of nature." "Nature is angry," another commented. and the video is AI-generated. A reverse image search led to more false posts, but Google's "About this image" feature identified the video as having been made with AI. The ability to detect AI-generated images is based on Google's SynthID technology, which was launched by its DeepMind AI lab in 2023. Image Screenshot of the image with a caption showing it is made with Google AI Several clear visual inconsistencies can also be seen in the clip -- telltale signs that it was made with AI. A person in white clothes sitting on the roof at the beginning of the video disappears later, while another man not visible at the beginning appears later in the video. Wooden "charpoy" beds also appear to be at the same height as buffaloes, when in reality they would be lower. Image Visual inconsistencies of the false video highlighted by AFP A reverse image search has also revealed the video was originally uploaded on July 3 to a TikTok account, which publishes other clips with similar visual inconsistencies (archived link). AFP has debunked other false claims sharing fabricated content created with AI here.
AFP
01-08-2025
- AFP
AI-generated videos used to tout non-existent Philippine government subsidy
"DSWD education cash assistance for everyone in school," reads the English and Tagalog-language caption of a Facebook video shared on July 9, 2025. The caption also includes a link where people can register for the purported cash assistance programme offered by the Philippines' Department of Social Welfare and Development (DSWD). The video shows a man wearing a DSWD branded shirt speaking to camera from a school yard, with students milling about in the background. He claims all students are entitled to varying amounts of educational cash assistance, depending on their grade and urges them to register online through the provided link. Superimposed text on the video says the cash will be paid out on July 12. The logos of the DSWD and the Department of Education are also superimposed. Image Screenshot of the false Facebook post captured on August 1, 2025, with a red X added by AFP The same video was also shared on TikTok, where it was viewed more than 1.2 million times. Similar videos were also shared elsewhere on TikTok and Facebook. These posts provide a link to a website where people can register their name, address and phone number, but the site is not a government webpage belonging to either the DSWD or education department. Both departments have also issued warnings about the posts. The DSWD said in a July 11 post on its verified Facebook page that the circulating posts and videos were "fake", reminding the public to be wary of information that does not come directly from its official social media accounts (archived link). An earlier post by the DSWD said it "does not offer direct financial assistance for education", but runs a scheme where students in need can receive assistance in exchange for tutoring elementary students who are struggling or have not yet learned to read (archived link). "AI videos being shared by a certain Facebook page about a supposed DSWD Educational Cash Assistance for all students is fake news," added an education department post on its verified Facebook page from July 17 (archived link). AI-generated videos A reverse image search led to more false posts, but Google's "About this image" feature identified the video as having been made with AI. The ability to detect AI-generated images is based on Google's SynthID technology, which was launched by its DeepMind AI lab in 2023 (archived here and here). Image Screenshot of the Google page indicating the video was made using AI A closer analysis of the video shows it contains a watermark in the bottom-right corner for Veo, Google's video-generation platform that allows users to create eight-second videos (archived here and here). The falsely shared video appears to have been created by splicing together three clips each around eight seconds long. Image Screenshot of the falsely shared video's final frames, with the Veo watermark magnified by AFP The falsely shared videos also contains visual inconsistencies, which despite the meteoric rise of AI technology remain an indicator of inauthentic visuals. In one video, for example, two students walking behind the speaker appear to merge into one. Image Screenshots showing two students in the falsely shared video appearing to merge, highlighted by AFP AFP has also previously debunked similar false claims about cash assistance programmes in the Philippines.
The Hindu
30-07-2025
- Business
- The Hindu
Google announces general availability for Veo 3 and Veo 3 Fast on Vertex AI
Google has made their AI video generation model, Veo 3 generally available on Vertex AI. The company also said Veo 3 Fast, the faster and more economical version of Veo 3, will also be generally available on Veo 3. The models will be open for public preview via Vertex AI in August and will offer the feature that generates video clips from images. Users just need to enter a text prompt along with the image. Google said that the videos generated would be watermarked with SynthID. Veo 3 and Veo Fast are also one of the tools covered by Google's indemnity for GenAI services. In a blog posted, the company shared that since Veo 3's preview launch on Vertex AI in June, enterprise clients had generated over 6 million videos signalling high demand.
Hans India
24-07-2025
- Entertainment
- Hans India
Google Photos Rolls Out Free AI Tools to Animate and Stylise Old Pictures
In a move that blends nostalgia with innovation, Google Photos has introduced a suite of free AI-powered tools designed to help users breathe new life into their favourite images. The update, announced on July 24, 2025, is part of Google's continued effort to bring artificial intelligence to the fingertips of everyday users, enriching how they interact with their digital memories. Now available in the United States on both Android and iOS devices, the latest features transform static photos into engaging short videos and stylised artworks. With over 1.5 billion users worldwide, Google Photos aims to democratize creative tools by making AI-enhanced features both accessible and intuitive. At the heart of the new capabilities is the photo-to-video function, powered by Google's advanced Veo 2 model. This tool allows users to select any photo and convert it into a dynamic six-second video. Depending on the mood or creative direction desired, users can choose prompts like 'Subtle movements' or 'I'm feeling lucky' to generate a unique visual effect. Similar features have been seen in YouTube and Gemini, but this integration brings that functionality directly into the photo gallery app, making it easier than ever to enhance personal content. Joining this feature is the Remix tool, which leverages the Imagen AI model to instantly stylise any photo. Whether users want to reimagine a moment in anime or give it a surreal 3D makeover, the tool makes it possible within seconds. These stylised outputs will be clearly marked with SynthID digital watermarks to ensure transparency about their AI-generated origins. Google is taking a user-first approach by labeling these features experimental and inviting community feedback. Users can provide input through a simple thumbs-up or thumbs-down mechanism, which Google says will help fine-tune the tools over time. The new features are part of a revamped "Create" tab within the Google Photos app, which consolidates all creative tools, both AI-powered and traditional, in one easy-to-navigate section. Alongside existing capabilities like collages and highlight videos, the addition of AI effects signals a broader push to turn the app into a playground for visual storytelling. Importantly, Google's expansion of AI within Photos reflects a larger trend across the company's product ecosystem. These updates were announced alongside similar enhancements to YouTube Shorts, which now also offers photo-to-video and AI-driven visual effects. By putting advanced generative models like Veo 2 and Imagen directly in the hands of users, Google continues its mission to enhance digital creativity while learning from real-world usage. As the company refines these tools through user feedback, it hopes to make the experience even more immersive and personalized. Whether you're looking to revisit cherished memories or explore artistic reinterpretations of your everyday snapshots, these new AI features offer a compelling reason to dive back into your photo library.
