Tech Roundup: Age Verification Rollout, Chrome Vertical Tabs & More

[A recurring feature on the latest in Science & Technology.]

• Ireland launches a test phase of its digital wallet including an age verification capability to keep teenagers safer online; uses a digital ID to facilitate age verification to help protect children and young people from harms on social media platforms; comes as Apple rolls out age verification to Singapore and South Korea.
• Google adds a search option in the Play Store's ratings and reviews section to help users find reviews matching certain keywords, but removes the ability to filter review by device model.
• Google quietly releases an offline-first dictation app called Google AI Edge Eloquent on iOS to take on Wispr Flow, SuperWhisper, Willow and others; also releases AI Edge Gallery for Android and iOS to run open-source AI models locally on mobile devices.
• Meta updates WhatsApp with support for Apple CarPlay, allowing users to more easily call and message their friends and family members through the app.
• Google officially brings support for vertical tabs in Chrome ands launches an improved reading mode for "immersive, text-focused reading experience."
• Google rolls out an AI Enhance button for Photos on Android globally, offering automated lighting and contrast adjustments and video playback speed controls.
• Elon Musk amends his OpenAI lawsuit to ask that damages he might win be awarded to OpenAI's charity arm and that Altman be removed from OpenAI's nonprofit board.
• Anthropic says Mythos Preview is a general-purpose model and found thousands of high-severity vulnerabilities, including some in every major OS and web browser; announces Project Glasswing, a cybersecurity initiative that will use its Claude Mythos Preview model to help find and fix software vulnerabilities.
• Indian government proposes new rules under which every social media user posting on current affairs can fall under the Ministry of Information and Broadcasting's (MIB) Code of Ethics regulatory framework, a statute now applicable only to registered news publishers.
• OpenAI says it's resetting Codex usage limits "to celebrate 3 million weekly codex users" and will reset them for every 1 million new users until it reaches 10 million.
• TikTok plans to invest €1 billion to build a second data center in Finland as part of its €12 billion European data sovereignty initiative for the data of more than 200 million users.
• Japan relaxes privacy laws, including removing the requirement for opt-in consent before sharing personal data, to accelerate the development and deployment of AI.
• Consumer advocacy group Public Interest Research Group (PIRG) finds laptops from Apple and Lenovo to be the least repairable; gives Apple a C-minus in laptop repairability and a D-minus in cell phone repairability.
• Mozilla accuses Microsoft of stacking the deck against its Firefox browser, arguing that design choices in Windows steer users toward Edge even when they explicitly choose another browser.
• A new study from AI Forensics finds that thousands of men are members of Telegram groups and channels that advertise and sell hacking and surveillance services that can be used to harass friends, wives and girlfriends, and former partners; says Telegram's design enables technology-facilitated gender-based violence (TFGBV) in Italy and Spain.
• Anthropic gets accused of degrading the performance of Claude Opus 4.6 and Claude Code, intentionally or as an outcome of compute limits, with users arguing that the company's flagship coding model feels less capable, less reliable and more wasteful with tokens.
• Russia's communications watchdog Roskomnadzor reportedly adds Bluesky to its registry of banned websites, as authorities continue to restrict access to foreign social media and messaging services, citing violations of Russian law or concerns about fraud.
• Anthropic requires government ID verification via Persona before subscription to Claude to "prevent abuse, enforce our usage policies and comply with legal obligations."
• Microsoft debuts new Copilot in Word capabilities for legal, finance and compliance professionals to help track changes and track changes when an audit trail is needed.
• New research from the Tech Transparency Project reveals that Apple and Google continue to host dozens of "nudify" despite policies prohibiting them; says searches for terms like "nudify," "undress," and "deepnude" in the app stores steers users to multiple apps capable of digitally stripping the clothes off women in photos, leading to 483 million collective downloads ands $122 million in lifetime revenue. (Both Apple and Google have removed many of the apps, with the latter noting that the International Age Rating Coalition, not Google, sets age ratings for apps in the Google Play Store.)
• The European Commission intends to order Meta to reinstate rival artificial intelligence assistants on its WhatsApp messaging service after the U.S. tech giant imposes an access fee, citing breach of E.U. competition rules.
• Italy's data protection regulator fines the national postal service provider Poste Italiane and Postepay SpA, a digital payments subsidiary, €12.5 million for data privacy violations by illegally processing millions of users' personal data; comes after the U.K.'s Information Commissioner's Office fined Imgur £247,590 for failing to use children’s personal information lawfully.
• Mozilla releases Firefox 150 with real-time, private translations via the about:translations page, along with the ability to reorder, copy, paste, delete and export pages in a PDF using the built-in PDF editor.
• Google brings AI Overviews to Gmail after launching it for Search and Drive; updates NotebookLM to let users auto-label and categorise sources (when there are five ore more items), and makes sharing easier by allowing users to pate an entire email list.
• GitHub says it has begun collecting pseudonymous client-side telemetry from command-line interface (CLI) users and enabled it by default, as Atlassian says it will begin collecting customer metadata and in-app content from Jira, Confluence and other cloud products by default on August 17, 2026, to train its AI offerings.
• OpenAI releases ChatGPT for Clinicians for medical tasks like documentation and research in the U.S.; makes it free for verified physicians and pharmacists; releases Privacy Filter, an open-weight model for masking personally identifiable information in text, with 1.5B total and 50M active parameters and announces workspace agents in ChatGPT, letting teams create Codex-powered shared agents for complex tasks.
• U.S. child safety group NCMEC says it received 1.5 million reports of suspected CSAM with ties to AI in 2025, a significant surge compared to 67,000 in 2024 and 4,700 in 2023.
• Alibaba launches Qwen3.6-27B, an open-weight dense model with 27B parameters; claims it surpasses Qwen3.5-397B-A17B on major coding benchmarks.
• Meta unveils Live Chats on Threads for real-time conversations during cultural events; plans to track employee use of popular sites and apps like Google and Wikipedia to train its AI models through an internal tool called the Model Capability Initiative (MCI), but insists there are safeguards in place to protect sensitive content, and the data is not used for any other purpose.
• Google says Meet's Gemini-powered Take Notes for Me feature can now be used for in-person meetings and adds support for Teams and Zoom; unveils Workspace Intelligence, which understands "complex semantic relationships" between data in Workspace apps to provide personalised context.
• Meta introduces an improved Meta Account as a centralised way to sign in and manage Meta apps and devices like Facebook, Instagram and AI glasses.