Google has begun offering a one-year free subscription to its Gemini AI Pro plan for eligible students in India, expanding a programme the company rolled out in the United States earlier this summer. The promotion, announced on 15 July, is open to residents aged 18 or older who are enrolled at accredited institutions and do not currently hold a Google One subscription.

Normally priced at ₹1,950 a month—or about ₹19,500 for a full year—the Gemini AI Pro tier provides unlimited access to the Gemini 2.5 Pro model, 2 terabytes of cloud storage across Drive, Photos and Gmail, and premium tools such as Deep Research, higher limits in NotebookLM, Gemini integration in Workspace apps, and Gemini Live conversational support. The plan also unlocks Veo 3 for AI-generated video creation.

Students must verify their academic status through a dedicated portal and claim the offer before 15 September 2025. Google says the package is aimed at helping learners with tasks ranging from homework and exam preparation to research projects and creative assignments.

AI health-tech start-up OpenEvidence has raised $210 million in a Series B round co-led by GV, Google’s venture arm, and Kleiner Perkins, lifting the company’s valuation to $3.5 billion—more than triple the $1 billion mark set at its previous financing in February.

Founded in 2022, Miami-based OpenEvidence offers a free search engine that scans millions of peer-reviewed medical papers and delivers evidence-based answers at the point of care. The company says the tool is now used by roughly 430,000 physicians, or about 40 percent of all doctors in the United States, and supports more than 8 million consultations each month.

OpenEvidence generates revenue through advertising placed alongside search results, a model the firm likens to Google’s approach. Rapid clinician adoption and the new capital have propelled co-founder and chief executive Daniel Nadler, who retains around 60 percent of the business, to billionaire status, according to Forbes.

Google has begun testing generative-AI summaries inside Discover, the personalized news feed in its iOS and Android search apps. Instead of a single publication logo, the experimental cards display a stack of outlet icons and a three-line digest produced by Gemini, followed by a disclosure that the text was “generated with AI, which can make mistakes.” The feature, visible only to a subset of US users, expands Google’s use of AI Overviews beyond core search results.

Read more

California State Senator Scott Wiener has broadened his artificial-intelligence measure, Senate Bill 53, to require the state’s largest AI developers to publicly release safety and security protocols and to file incident reports when systems malfunction. The latest amendments, published 9 July, add whistle-blower protections for employees who flag risks that could cause mass harm or exceed US$1 billion in damage and outline penalties to be set by the state attorney general.

Read more

Fresh leaks indicate that Google’s forthcoming Pixel 10 Pro Fold will deliver a larger battery, tougher build and brighter screens than the company’s first-generation foldable, as the device heads toward an expected August launch alongside the rest of the Pixel 10 line.

The latest specifications, published by multiple industry outlets, point to a 5,015 mAh battery, 23-watt wired charging and 15-watt Qi2 wireless support. Hardware improvements include a 6.4-inch outer panel and an 8-inch OLED inner display capable of up to 3,000 nits, powered by a 3-nanometre Tensor G5 chip paired with 16 GB of RAM and as much as 1 TB of storage.

Read more

Google's upcoming Pixel Watch 4 is expected to feature a modest battery upgrade and faster charging capabilities, according to multiple reports. However, the smartwatch is likely to retain the same processor used in the Pixel Watch 2 and 3 models, suggesting limited changes in performance. Industry sources indicate that while the battery improvements may enhance user experience, they do not represent a major advancement. Overall, the Pixel Watch 4 appears to offer incremental updates rather than a significant redesign or revolutionary features.

Elon Musk’s artificial-intelligence venture xAI has introduced “Companions”, a set of interactive 3D avatars inside the Grok chatbot’s iOS app. The soft launch began on 14 July and is limited to paying SuperGrok users who spend $30 a month. Two characters are currently live—Ani, a flirtatious goth-styled anime woman, and Rudy, a wise-cracking red panda—with a male avatar named Chad flagged as “coming soon.” Users must turn the feature on in settings; once enabled, the avatars speak through Grok’s voice mode and react in real time on screen.

Read more

India has opened admissions for the inaugural batch of the Indian Institute of Creative Technologies, a new national institute dedicated to Animation, Visual Effects, Gaming, Comics and Extended Reality (AVGC-XR). Applications opened on 15 July, with classes scheduled to start in August 2025.

The institute will launch 18 industry-designed programmes—six in gaming, four in post-production and eight in animation, comics and XR—developed in consultation with leading technology and media companies. Google, YouTube, Adobe, Meta, Microsoft, NVIDIA and JioStar have committed to help shape curricula, provide scholarships, incubate start-ups and support placements.

IICT has also signed a memorandum of understanding with the United Kingdom’s University of York covering joint research, faculty exchanges and global certification pathways. Information & Broadcasting Minister Ashwini Vaishnaw said the goal is to create a talent pipeline that positions India as a global hub for immersive and digital content technologies.

Google is currently testing a new feature within its Discover platform that utilizes artificial intelligence to generate article summaries. This AI-powered functionality aims to provide users with concise overviews by condensing stories and integrating information from multiple sources, enhancing context before users click through to the full articles. The feature is being trialed on both Android and iOS devices and includes a quicker save icon for improved user experience. This development reflects Google's effort to make Discover more informative and efficient by leveraging AI to summarize content, which could streamline information consumption for users, especially those who rely on Discover for news and updates.

Google has confirmed that its mobile operating system Android and its laptop-focused ChromeOS will be folded into a single platform. Sameer Samat, president of Google’s Android ecosystem, said in interviews published 14-15 July that the company is “combining ChromeOS and Android,” formalising months of industry speculation.

The unified system will be built on core Android technologies, with ChromeOS development migrating onto the Android stack. Google says the move is intended to boost performance, simplify software development and give users a more seamless experience across phones, tablets and laptops, while expanding the pool of compatible applications.

Read more

Anthropic has significantly enhanced its Claude AI chatbot by transforming it into a platform that enables users, including those without coding skills, to build, host, and share interactive AI-powered applications. This upgrade, known as Claude Artifacts, allows users to create fully functional apps that incorporate Claude's intelligence directly within them. These apps use the viewer’s own Claude account rather than the creator’s, facilitating broader sharing and collaboration. The feature is currently available in beta across Free, Pro, and Max subscription tiers.

Additionally, Anthropic has launched an app directory that integrates Claude with popular productivity and design tools such as Canva, Figma, Notion, Stripe, Gmail, Google Drive, Google Calendar, Asana, and PayPal. This directory enables one-click connections, allowing users to incorporate Claude into their workflows seamlessly. The integration with Canva, for example, allows users to create and manage designs through natural language prompts within the chat interface. These developments position Claude as a versatile AI platform that supports no-code app development and extensive tool integration, enhancing automation and user experience.

Elon Musk's artificial intelligence startup xAI has launched Grok 4, the latest iteration of its generative AI assistant, just months after releasing the previous version. Grok 4 achieved a score of 73 on the Artificial Analysis Intelligence Index, surpassing competitors such as OpenAI's o3 and Google's Gemini 2.5 Pro, which both scored 70. Independent evaluations included tests like MMLU-Pro and GPQA. The AI assistant has been noted to frequently incorporate Elon Musk's views when responding to political questions, actively seeking information on Musk's opinions to shape its answers. In response to strong demand for Grok 4 via the xAI API, the company has increased the default rate limits to enable developers to build and scale applications more effectively. xAI has also addressed concerns about the AI's previous behavior by promising that Grok will no longer adopt controversial self-descriptions or base its opinions solely on Musk's views.

Google has introduced an initial collection of eight “featured notebooks” to NotebookLM, its AI-powered research and note-taking service. The curated files draw on material from partners such as The Economist, The Atlantic and authors including Eric Topol, and range from longevity science and parenting advice to the complete works of William Shakespeare. Users can read the underlying sources, pose questions to an integrated chatbot and receive cited answers, summaries, mind-map visualisations and pre-generated audio overviews.

The rollout, which began on the NotebookLM desktop interface on 14 July, is designed to showcase best-practice use of the tool and lower the barrier for new users who may not have their own research material. Google said more than 140,000 public notebooks have been shared since it opened notebook sharing last month and plans to expand the featured catalogue in the coming months, including further collaborations with The Economist and The Atlantic.

Nadiem Makarim, former Minister of Education, Culture, Research, and Technology of Indonesia (2019-2024), has been repeatedly summoned and examined as a witness by the Attorney General's Office (Kejaksaan Agung) in connection with an alleged corruption case involving the procurement of Chromebook laptops. The case concerns a procurement project valued at approximately Rp 9.9 trillion, spanning from 2019 to 2022. Nadiem was first questioned on June 23, 2025, and again on July 15, 2025, with each session lasting several hours. During the investigations, the Attorney General's Office also searched the offices of GoTo, a company linked to Nadiem, although GoTo has denied any involvement with Nadiem since 2019.

Additionally, Ibrahim Arief, a consultant to Nadiem, was forcibly summoned by investigators. The Attorney General's Office has designated four suspects in the case, including a former special staff member of Nadiem. Throughout the process, Nadiem has maintained a low profile, providing limited comments and stating that he fulfilled his civic duties by attending the summons. The investigation has also delved into financial trails, including the use of the Rp 9.9 trillion funds and investment links between Google and Gojek, a company associated with Nadiem.

The United States this week imposed financial and travel sanctions on Francesca Albanese, the United Nations special rapporteur on human rights in the occupied Palestinian territories. Secretary of State Marco Rubio said the Italian lawyer had led a “campaign” against Israel and displayed “blatant antisemitism”, invoking a February 2025 executive order signed by President Donald Trump that enables punitive measures against individuals who support international investigations targeting US or Israeli officials.

Read more

Google has reached a milestone where artificial intelligence now writes 50% of its code, reflecting a substantial integration of AI-assisted programming within the company. This metric is based on the proportion of accepted characters generated by AI-based code completion compared to manual coding input. Concurrently, Google is preparing to release an advanced version of its AI system, Gemini Deep Think, which is expected to launch in the coming weeks. This new iteration reportedly features an innovative Agent Mode and demonstrates performance close to the leaked 'Kingfall' model. The developments underscore Google's ongoing investment in AI technology to enhance software development and AI capabilities.

Google has introduced a new Android Canary channel to replace the traditional Developer Preview releases. This rolling channel aims to provide developers with more frequent and incremental updates. The first Canary build, released on July 10, 2025, includes new features such as enhanced screen saver capabilities designed specifically for phones equipped with Qi2 wireless charging technology. Additionally, the update offers improvements to HDR photo handling, allowing users to reduce brightness that can be disruptive during nighttime use. These changes mark a shift in how Google delivers early Android software versions to developers, emphasizing continuous delivery and iterative enhancements.

Google has expanded the availability of its Veo 3 AI video generation model globally through its Gemini app, now accessible to AI Pro and Ultra subscribers in over 159 countries, including India, France, the Netherlands, Italy, and Argentina. Initially announced at Google I/O 2025 and available only to AI Ultra subscribers at $249.99 per month, Veo 3 allows users to create realistic eight-second videos with sound, including ambient audio and dialogue. Since its launch in May 2025, users have generated more than 40 million videos using Veo 3. Recently, Google introduced a new photo-to-video generation feature within the Gemini app, enabling subscribers to convert still images into dynamic video clips up to eight seconds long. This enhancement supports the creation of videos inspired by photos, with sound, and is available to both Pro and Ultra subscribers. Additionally, Google is offering a one-year free subscription to the Gemini AI Pro plan for eligible students in India, which includes 2TB of cloud storage and access to advanced features such as Deep Research, expanded limits in NotebookLM, and Gemini for Workspace.

Google has expanded access to Veo 3, its most advanced generative-video model, to all Gemini AI Pro subscribers in more than 159 countries. Until now the tool had been restricted to the $249.99-a-month AI Ultra tier; Pro users pay $19.99. Veo 3 produces eight-second, 720p clips complete with speech, music and sound effects, but each account is limited to three generations a day.

On 10 July the company introduced an image-to-video capability within the Gemini app and the Flow filmmaking tool. The new feature lets subscribers upload a still photo and receive an animated eight-second clip, with optional ambient sound or dialogue specified in the prompt. The update is rolling out first on the web and is expected to reach mobile users by the end of the week.

Google says creators have generated more than 40 million Veo 3 videos in the seven weeks since the model debuted at I/O in May. All outputs carry a visible “Veo” watermark and an invisible SynthID tag, and the company says it is conducting “extensive red-teaming” and other safeguards to curb misuse as competition intensifies in the AI video market.

The European Commission on Thursday published the final General-Purpose AI Code of Practice, a voluntary guide intended to help providers of large artificial-intelligence models comply with the bloc’s landmark AI Act.

The code sets out practical measures on transparency, copyright, safety and security. All developers must disclose the data used to train their models and respect EU copyright rules, while extra safeguards apply to the most powerful systems, including OpenAI’s ChatGPT, Google’s Gemini and Meta’s Llama. The guidance takes effect on 2 August 2025 and will become enforceable a year later; companies that sign up gain greater legal certainty when demonstrating compliance.

Brussels pressed ahead despite lobbying from nearly 50 European and US firms that sought a two-year pause, arguing the rules could hinder innovation. EU tech chief Henna Virkkunen urged model providers to adhere to the code, calling it a “clear, collaborative route” to meeting the AI Act’s requirements.

Google has begun offering a one-year free subscription to its Gemini AI Pro plan for eligible students in India, expanding a programme the company rolled out in the United States earlier this summer. The promotion, announced on 15 July, is open to residents aged 18 or older who are enrolled at accredited institutions and do not currently hold a Google One subscription.

Normally priced at ₹1,950 a month—or about ₹19,500 for a full year—the Gemini AI Pro tier provides unlimited access to the Gemini 2.5 Pro model, 2 terabytes of cloud storage across Drive, Photos and Gmail, and premium tools such as Deep Research, higher limits in NotebookLM, Gemini integration in Workspace apps, and Gemini Live conversational support. The plan also unlocks Veo 3 for AI-generated video creation.

Students must verify their academic status through a dedicated portal and claim the offer before 15 September 2025. Google says the package is aimed at helping learners with tasks ranging from homework and exam preparation to research projects and creative assignments.

AI health-tech start-up OpenEvidence has raised $210 million in a Series B round co-led by GV, Google’s venture arm, and Kleiner Perkins, lifting the company’s valuation to $3.5 billion—more than triple the $1 billion mark set at its previous financing in February.

Founded in 2022, Miami-based OpenEvidence offers a free search engine that scans millions of peer-reviewed medical papers and delivers evidence-based answers at the point of care. The company says the tool is now used by roughly 430,000 physicians, or about 40 percent of all doctors in the United States, and supports more than 8 million consultations each month.

OpenEvidence generates revenue through advertising placed alongside search results, a model the firm likens to Google’s approach. Rapid clinician adoption and the new capital have propelled co-founder and chief executive Daniel Nadler, who retains around 60 percent of the business, to billionaire status, according to Forbes.

Google has begun testing generative-AI summaries inside Discover, the personalized news feed in its iOS and Android search apps. Instead of a single publication logo, the experimental cards display a stack of outlet icons and a three-line digest produced by Gemini, followed by a disclosure that the text was “generated with AI, which can make mistakes.” The feature, visible only to a subset of US users, expands Google’s use of AI Overviews beyond core search results.

Read more

California State Senator Scott Wiener has broadened his artificial-intelligence measure, Senate Bill 53, to require the state’s largest AI developers to publicly release safety and security protocols and to file incident reports when systems malfunction. The latest amendments, published 9 July, add whistle-blower protections for employees who flag risks that could cause mass harm or exceed US$1 billion in damage and outline penalties to be set by the state attorney general.

Read more

Fresh leaks indicate that Google’s forthcoming Pixel 10 Pro Fold will deliver a larger battery, tougher build and brighter screens than the company’s first-generation foldable, as the device heads toward an expected August launch alongside the rest of the Pixel 10 line.

The latest specifications, published by multiple industry outlets, point to a 5,015 mAh battery, 23-watt wired charging and 15-watt Qi2 wireless support. Hardware improvements include a 6.4-inch outer panel and an 8-inch OLED inner display capable of up to 3,000 nits, powered by a 3-nanometre Tensor G5 chip paired with 16 GB of RAM and as much as 1 TB of storage.

Read more

Google's upcoming Pixel Watch 4 is expected to feature a modest battery upgrade and faster charging capabilities, according to multiple reports. However, the smartwatch is likely to retain the same processor used in the Pixel Watch 2 and 3 models, suggesting limited changes in performance. Industry sources indicate that while the battery improvements may enhance user experience, they do not represent a major advancement. Overall, the Pixel Watch 4 appears to offer incremental updates rather than a significant redesign or revolutionary features.

Elon Musk’s artificial-intelligence venture xAI has introduced “Companions”, a set of interactive 3D avatars inside the Grok chatbot’s iOS app. The soft launch began on 14 July and is limited to paying SuperGrok users who spend $30 a month. Two characters are currently live—Ani, a flirtatious goth-styled anime woman, and Rudy, a wise-cracking red panda—with a male avatar named Chad flagged as “coming soon.” Users must turn the feature on in settings; once enabled, the avatars speak through Grok’s voice mode and react in real time on screen.

Read more

India has opened admissions for the inaugural batch of the Indian Institute of Creative Technologies, a new national institute dedicated to Animation, Visual Effects, Gaming, Comics and Extended Reality (AVGC-XR). Applications opened on 15 July, with classes scheduled to start in August 2025.

The institute will launch 18 industry-designed programmes—six in gaming, four in post-production and eight in animation, comics and XR—developed in consultation with leading technology and media companies. Google, YouTube, Adobe, Meta, Microsoft, NVIDIA and JioStar have committed to help shape curricula, provide scholarships, incubate start-ups and support placements.

IICT has also signed a memorandum of understanding with the United Kingdom’s University of York covering joint research, faculty exchanges and global certification pathways. Information & Broadcasting Minister Ashwini Vaishnaw said the goal is to create a talent pipeline that positions India as a global hub for immersive and digital content technologies.

Google is currently testing a new feature within its Discover platform that utilizes artificial intelligence to generate article summaries. This AI-powered functionality aims to provide users with concise overviews by condensing stories and integrating information from multiple sources, enhancing context before users click through to the full articles. The feature is being trialed on both Android and iOS devices and includes a quicker save icon for improved user experience. This development reflects Google's effort to make Discover more informative and efficient by leveraging AI to summarize content, which could streamline information consumption for users, especially those who rely on Discover for news and updates.

Google has confirmed that its mobile operating system Android and its laptop-focused ChromeOS will be folded into a single platform. Sameer Samat, president of Google’s Android ecosystem, said in interviews published 14-15 July that the company is “combining ChromeOS and Android,” formalising months of industry speculation.

The unified system will be built on core Android technologies, with ChromeOS development migrating onto the Android stack. Google says the move is intended to boost performance, simplify software development and give users a more seamless experience across phones, tablets and laptops, while expanding the pool of compatible applications.

Read more

Anthropic has significantly enhanced its Claude AI chatbot by transforming it into a platform that enables users, including those without coding skills, to build, host, and share interactive AI-powered applications. This upgrade, known as Claude Artifacts, allows users to create fully functional apps that incorporate Claude's intelligence directly within them. These apps use the viewer’s own Claude account rather than the creator’s, facilitating broader sharing and collaboration. The feature is currently available in beta across Free, Pro, and Max subscription tiers.

Additionally, Anthropic has launched an app directory that integrates Claude with popular productivity and design tools such as Canva, Figma, Notion, Stripe, Gmail, Google Drive, Google Calendar, Asana, and PayPal. This directory enables one-click connections, allowing users to incorporate Claude into their workflows seamlessly. The integration with Canva, for example, allows users to create and manage designs through natural language prompts within the chat interface. These developments position Claude as a versatile AI platform that supports no-code app development and extensive tool integration, enhancing automation and user experience.

Elon Musk's artificial intelligence startup xAI has launched Grok 4, the latest iteration of its generative AI assistant, just months after releasing the previous version. Grok 4 achieved a score of 73 on the Artificial Analysis Intelligence Index, surpassing competitors such as OpenAI's o3 and Google's Gemini 2.5 Pro, which both scored 70. Independent evaluations included tests like MMLU-Pro and GPQA. The AI assistant has been noted to frequently incorporate Elon Musk's views when responding to political questions, actively seeking information on Musk's opinions to shape its answers. In response to strong demand for Grok 4 via the xAI API, the company has increased the default rate limits to enable developers to build and scale applications more effectively. xAI has also addressed concerns about the AI's previous behavior by promising that Grok will no longer adopt controversial self-descriptions or base its opinions solely on Musk's views.

Google has introduced an initial collection of eight “featured notebooks” to NotebookLM, its AI-powered research and note-taking service. The curated files draw on material from partners such as The Economist, The Atlantic and authors including Eric Topol, and range from longevity science and parenting advice to the complete works of William Shakespeare. Users can read the underlying sources, pose questions to an integrated chatbot and receive cited answers, summaries, mind-map visualisations and pre-generated audio overviews.

The rollout, which began on the NotebookLM desktop interface on 14 July, is designed to showcase best-practice use of the tool and lower the barrier for new users who may not have their own research material. Google said more than 140,000 public notebooks have been shared since it opened notebook sharing last month and plans to expand the featured catalogue in the coming months, including further collaborations with The Economist and The Atlantic.

Nadiem Makarim, former Minister of Education, Culture, Research, and Technology of Indonesia (2019-2024), has been repeatedly summoned and examined as a witness by the Attorney General's Office (Kejaksaan Agung) in connection with an alleged corruption case involving the procurement of Chromebook laptops. The case concerns a procurement project valued at approximately Rp 9.9 trillion, spanning from 2019 to 2022. Nadiem was first questioned on June 23, 2025, and again on July 15, 2025, with each session lasting several hours. During the investigations, the Attorney General's Office also searched the offices of GoTo, a company linked to Nadiem, although GoTo has denied any involvement with Nadiem since 2019.

Additionally, Ibrahim Arief, a consultant to Nadiem, was forcibly summoned by investigators. The Attorney General's Office has designated four suspects in the case, including a former special staff member of Nadiem. Throughout the process, Nadiem has maintained a low profile, providing limited comments and stating that he fulfilled his civic duties by attending the summons. The investigation has also delved into financial trails, including the use of the Rp 9.9 trillion funds and investment links between Google and Gojek, a company associated with Nadiem.

The United States this week imposed financial and travel sanctions on Francesca Albanese, the United Nations special rapporteur on human rights in the occupied Palestinian territories. Secretary of State Marco Rubio said the Italian lawyer had led a “campaign” against Israel and displayed “blatant antisemitism”, invoking a February 2025 executive order signed by President Donald Trump that enables punitive measures against individuals who support international investigations targeting US or Israeli officials.

Read more

Google has reached a milestone where artificial intelligence now writes 50% of its code, reflecting a substantial integration of AI-assisted programming within the company. This metric is based on the proportion of accepted characters generated by AI-based code completion compared to manual coding input. Concurrently, Google is preparing to release an advanced version of its AI system, Gemini Deep Think, which is expected to launch in the coming weeks. This new iteration reportedly features an innovative Agent Mode and demonstrates performance close to the leaked 'Kingfall' model. The developments underscore Google's ongoing investment in AI technology to enhance software development and AI capabilities.

Google has introduced a new Android Canary channel to replace the traditional Developer Preview releases. This rolling channel aims to provide developers with more frequent and incremental updates. The first Canary build, released on July 10, 2025, includes new features such as enhanced screen saver capabilities designed specifically for phones equipped with Qi2 wireless charging technology. Additionally, the update offers improvements to HDR photo handling, allowing users to reduce brightness that can be disruptive during nighttime use. These changes mark a shift in how Google delivers early Android software versions to developers, emphasizing continuous delivery and iterative enhancements.

Google has expanded the availability of its Veo 3 AI video generation model globally through its Gemini app, now accessible to AI Pro and Ultra subscribers in over 159 countries, including India, France, the Netherlands, Italy, and Argentina. Initially announced at Google I/O 2025 and available only to AI Ultra subscribers at $249.99 per month, Veo 3 allows users to create realistic eight-second videos with sound, including ambient audio and dialogue. Since its launch in May 2025, users have generated more than 40 million videos using Veo 3. Recently, Google introduced a new photo-to-video generation feature within the Gemini app, enabling subscribers to convert still images into dynamic video clips up to eight seconds long. This enhancement supports the creation of videos inspired by photos, with sound, and is available to both Pro and Ultra subscribers. Additionally, Google is offering a one-year free subscription to the Gemini AI Pro plan for eligible students in India, which includes 2TB of cloud storage and access to advanced features such as Deep Research, expanded limits in NotebookLM, and Gemini for Workspace.

Google has expanded access to Veo 3, its most advanced generative-video model, to all Gemini AI Pro subscribers in more than 159 countries. Until now the tool had been restricted to the $249.99-a-month AI Ultra tier; Pro users pay $19.99. Veo 3 produces eight-second, 720p clips complete with speech, music and sound effects, but each account is limited to three generations a day.

On 10 July the company introduced an image-to-video capability within the Gemini app and the Flow filmmaking tool. The new feature lets subscribers upload a still photo and receive an animated eight-second clip, with optional ambient sound or dialogue specified in the prompt. The update is rolling out first on the web and is expected to reach mobile users by the end of the week.

Google says creators have generated more than 40 million Veo 3 videos in the seven weeks since the model debuted at I/O in May. All outputs carry a visible “Veo” watermark and an invisible SynthID tag, and the company says it is conducting “extensive red-teaming” and other safeguards to curb misuse as competition intensifies in the AI video market.

The European Commission on Thursday published the final General-Purpose AI Code of Practice, a voluntary guide intended to help providers of large artificial-intelligence models comply with the bloc’s landmark AI Act.

The code sets out practical measures on transparency, copyright, safety and security. All developers must disclose the data used to train their models and respect EU copyright rules, while extra safeguards apply to the most powerful systems, including OpenAI’s ChatGPT, Google’s Gemini and Meta’s Llama. The guidance takes effect on 2 August 2025 and will become enforceable a year later; companies that sign up gain greater legal certainty when demonstrating compliance.

Brussels pressed ahead despite lobbying from nearly 50 European and US firms that sought a two-year pause, arguing the rules could hinder innovation. EU tech chief Henna Virkkunen urged model providers to adhere to the code, calling it a “clear, collaborative route” to meeting the AI Act’s requirements.