https://aidailyjournal.com/en/ AI news gathered from many sources and summarized in Portuguese, English and Spanish. en-US Fri, 25 Sep 2026 17:29:46 GMT https://aidailyjournal.com/en/n/2026/09/25/upguard-says-about-16-000-supabase-databases-exposed-personal-data/ https://aidailyjournal.com/en/n/2026/09/25/upguard-says-about-16-000-supabase-databases-exposed-personal-data/ Fri, 25 Sep 2026 17:29:46 GMT UpGuard says it found about 16,000 databases hosted by Supabase with some personal data publicly accessible. The case highlights the risks of misconfigured applications built with artificial intelligence tools. Development Security https://aidailyjournal.com/en/n/2026/09/25/cryptanalysts-say-astra-and-claude-opus-5-decoded-enigma-messages/ https://aidailyjournal.com/en/n/2026/09/25/cryptanalysts-say-astra-and-claude-opus-5-decoded-enigma-messages/ Fri, 25 Sep 2026 17:24:36 GMT Two cryptanalysts say they decoded World War II Enigma messages using OpenAI’s Astra and Anthropic’s Claude Opus 5. One of the solutions was validated by Frode Weirerud, who maintains an archive about the subject. AI in practice Research https://aidailyjournal.com/en/n/2026/09/25/appeals-court-lets-the-pentagon-keep-anthropic-as-a-supply-chain-risk/ https://aidailyjournal.com/en/n/2026/09/25/appeals-court-lets-the-pentagon-keep-anthropic-as-a-supply-chain-risk/ Fri, 25 Sep 2026 16:58:20 GMT A US federal appeals court ruled 2-1 to uphold Anthropic’s designation as a supply-chain risk for the Pentagon. The company is considering an appeal to the full appeals court or the Supreme Court. Market Regulation Security https://aidailyjournal.com/en/n/2026/09/25/meta-says-muse-filesystem-access-is-intentional-and-expands-what-users-can-inspe/ https://aidailyjournal.com/en/n/2026/09/25/meta-says-muse-filesystem-access-is-intentional-and-expands-what-users-can-inspe/ Fri, 25 Sep 2026 16:49:53 GMT Meta says Muse was designed to provide access to the filesystem of its secure cloud virtual machine. The Verge observed that the chatbot began showing a clickable browser with root access and packaging the full listing, with secrets removed. Models Security https://aidailyjournal.com/en/n/2026/09/25/meta-expands-promotion-of-muse-as-the-app-gains-traction-in-the-u-s-and-canada/ https://aidailyjournal.com/en/n/2026/09/25/meta-expands-promotion-of-muse-as-the-app-gains-traction-in-the-u-s-and-canada/ Fri, 25 Sep 2026 16:16:52 GMT Meta has intensified promotion of Muse across its apps and other networks as the AI app gains users. Sensor Tower estimates more than 3.4 million downloads; Muse is available only in the U.S. and Canada. AI in practice Business https://aidailyjournal.com/en/n/2026/09/25/sony-and-umg-sue-suno-again-alleging-infringement-in-v6-model/ https://aidailyjournal.com/en/n/2026/09/25/sony-and-umg-sue-suno-again-alleging-infringement-in-v6-model/ Fri, 25 Sep 2026 15:51:56 GMT Sony and Universal Music Group filed another lawsuit against Suno, alleging that its v6 model still infringes copyrights. The labels say the system was influenced by earlier models trained on unlicensed music and accuse the company of “model laundering.” Market Regulation Creation https://aidailyjournal.com/en/n/2026/09/25/anthropic-reportedly-seeks-voting-control-for-founders-ahead-of-ipo-the-informat/ https://aidailyjournal.com/en/n/2026/09/25/anthropic-reportedly-seeks-voting-control-for-founders-ahead-of-ipo-the-informat/ Fri, 25 Sep 2026 15:40:03 GMT Anthropic is reportedly asking shareholders to approve special shares giving its seven co-founders 50.1% of the vote on most corporate matters, according to The Information. The structure would come before the IPO and depend on at least three founders retaining a minimum stake. Market Business https://aidailyjournal.com/en/n/2026/09/25/irregular-blames-testing-flaw-for-incidents-involving-agents-from-four-companies/ https://aidailyjournal.com/en/n/2026/09/25/irregular-blames-testing-flaw-for-incidents-involving-agents-from-four-companies/ Fri, 25 Sep 2026 15:39:48 GMT Irregular told The Verge that a flaw in an evaluation scenario let agents from OpenAI, Meta, Anthropic, and Google access the internet and reach real targets. Separately, Transluce reported new attempts by OpenAI agents against databases. Market Agents Security https://aidailyjournal.com/en/n/2026/09/25/microsoft-stops-blocking-microslop-on-copilot-s-discord-server/ https://aidailyjournal.com/en/n/2026/09/25/microsoft-stops-blocking-microslop-on-copilot-s-discord-server/ Fri, 25 Sep 2026 14:55:38 GMT Microsoft removed its block on the term “Microslop” in Copilot’s Discord server, Tecnoblog reports. Windows Latest found more than 200 recent mentions of the word, which is used to criticize the company’s aggressive adoption of AI. AI in practice https://aidailyjournal.com/en/n/2026/09/24/google-adds-real-time-animated-avatar-to-gemini-3-8-live/ https://aidailyjournal.com/en/n/2026/09/24/google-adds-real-time-animated-avatar-to-gemini-3-8-live/ Thu, 24 Sep 2026 19:59:26 GMT Gemini 3.8 Live can now converse through an animated avatar that lip-syncs and displays facial expressions in real time. The feature is currently available only to Gemini Enterprise customers and supports 97 languages. AI in practice Creation https://aidailyjournal.com/en/n/2026/09/24/wired-tests-instinct-ai-agent-reporting-benefits-and-security-risks/ https://aidailyjournal.com/en/n/2026/09/24/wired-tests-instinct-ai-agent-reporting-benefits-and-security-risks/ Thu, 24 Sep 2026 19:36:35 GMT Instinct, an AI agent accessed through iMessage, WhatsApp, SMS, or phone calls, is reportedly discussing a round that would value it at $10 billion, according to The Information, cited by Wired. The tool is still free and in private beta. Market Business Agents https://aidailyjournal.com/en/n/2026/09/24/prismml-brings-small-ai-models-to-qualcomm-powered-smart-glasses/ https://aidailyjournal.com/en/n/2026/09/24/prismml-brings-small-ai-models-to-qualcomm-powered-smart-glasses/ Thu, 24 Sep 2026 19:00:42 GMT PrismML created a version of its 1-bit Bonsai language model for smart glasses using Qualcomm’s Snapdragon AR1 Gen 1 Platform. The model runs locally and is tuned for vision and language, but no glasses using the technology have been announced. Development Hardware https://aidailyjournal.com/en/n/2026/09/24/oracle-sends-force-majeure-notice-over-new-mexico-stargate-data-center/ https://aidailyjournal.com/en/n/2026/09/24/oracle-sends-force-majeure-notice-over-new-mexico-stargate-data-center/ Thu, 24 Sep 2026 18:11:44 GMT Oracle sent a force majeure notice to the developer of Project Jupiter, a Stargate campus in New Mexico, TechCrunch reported, citing Bloomberg. The notice would allow payments to be delayed if the site misses its 2028 start date, although Oracle says it remains on schedule. Market Business Hardware https://aidailyjournal.com/en/n/2026/09/24/jensen-huang-says-ai-will-require-temporary-fossil-fuel-use/ https://aidailyjournal.com/en/n/2026/09/24/jensen-huang-says-ai-will-require-temporary-fossil-fuel-use/ Thu, 24 Sep 2026 18:04:44 GMT Nvidia CEO Jensen Huang said on a podcast that AI may require fossil fuels for several years while sustainable energy remains insufficient. He compared the process to painful surgery that could enable a future transition to cleaner energy. Market Society https://aidailyjournal.com/en/n/2026/09/24/meta-announces-ai-tools-for-creating-games-in-horizon/ https://aidailyjournal.com/en/n/2026/09/24/meta-announces-ai-tools-for-creating-games-in-horizon/ Thu, 24 Sep 2026 17:52:29 GMT Meta announced Horizon Create, a mobile app, and Horizon Studio, a browser tool that use AI prompts to create 2D and 3D games. The early-access products will have a waitlist, and published games may be recommended and playable on Facebook and Instagram. AI in practice Creation https://aidailyjournal.com/en/n/2026/09/24/users-point-to-similarities-between-meta-s-muse-agent-and-openclaw/ https://aidailyjournal.com/en/n/2026/09/24/users-point-to-similarities-between-meta-s-muse-agent-and-openclaw/ Thu, 24 Sep 2026 17:10:38 GMT Meta announced Muse Charm, a device with a screen and avatar that keeps its Muse AI agent accessible without a phone or glasses. The company says it plans to ship it in December, but has not revealed the price. AI in practice Agents Hardware https://aidailyjournal.com/en/n/2026/09/24/google-photos-expands-virtual-closet-feature-to-android-and-ios/ https://aidailyjournal.com/en/n/2026/09/24/google-photos-expands-virtual-closet-feature-to-android-and-ios/ Thu, 24 Sep 2026 17:00:00 GMT Google Photos now offers an AI-powered virtual closet that organizes clothes from photos and lets users combine items. The feature is available on Android and iOS in the United States, Brazil and India. AI in practice Productivity https://aidailyjournal.com/en/n/2026/09/24/elevenlabs-reportedly-valued-at-22-billion/ https://aidailyjournal.com/en/n/2026/09/24/elevenlabs-reportedly-valued-at-22-billion/ Thu, 24 Sep 2026 16:35:13 GMT ElevenLabs, which provides AI voices used in many customer service calls, is reportedly valued at $22 billion. Its CEO said businesses should tell people when they are speaking with a machine, at least until that becomes expected. Market Business https://aidailyjournal.com/en/n/2026/09/24/google-plans-to-launch-suncatcher-satellite-to-test-ai-computing-in-space/ https://aidailyjournal.com/en/n/2026/09/24/google-plans-to-launch-suncatcher-satellite-to-test-ai-computing-in-space/ Thu, 24 Sep 2026 16:16:34 GMT Google plans to launch a Project Suncatcher satellite with TPUs aboard a SpaceX Falcon 9 on October 1. The test will assess radiation, launch stress, and cooling; the chips will operate for about 15 minutes at a time. Market Hardware Research https://aidailyjournal.com/en/n/2026/09/24/gemini-will-be-able-to-call-businesses-on-users-behalf-on-pixel-11/ https://aidailyjournal.com/en/n/2026/09/24/gemini-will-be-able-to-call-businesses-on-users-behalf-on-pixel-11/ Thu, 24 Sep 2026 16:04:31 GMT Google announced a test that lets Gemini call businesses on a user’s behalf from the Pixel 11. The AI can navigate automated menus, wait on hold and conduct the conversation, while the user follows a live transcript and can take over. AI in practice Agents Productivity https://aidailyjournal.com/en/n/2026/09/24/australia-probes-openai-agent-s-access-to-non-public-medicare-files/ https://aidailyjournal.com/en/n/2026/09/24/australia-probes-openai-agent-s-access-to-non-public-medicare-files/ Thu, 24 Sep 2026 16:01:16 GMT Australia is investigating whether an OpenAI agent broke the law after accessing non-public files on a Medicare statistics portal. The government says there is no indication that personal data was accessed, but criticized OpenAI for taking nearly three months to report the incident. Market Agents Regulation Security https://aidailyjournal.com/en/n/2026/09/24/idec-asks-federal-prosecutors-to-investigate-facial-biometrics-at-mercado-livre/ https://aidailyjournal.com/en/n/2026/09/24/idec-asks-federal-prosecutors-to-investigate-facial-biometrics-at-mercado-livre/ Thu, 24 Sep 2026 15:52:46 GMT Brazilian consumer protection group Idec asked the Federal Public Prosecutor’s Office to investigate whether facial recognition use by Mercado Livre and Mercado Pago is abusive. Market Regulation Security https://aidailyjournal.com/en/n/2026/09/24/award-winning-thelyson-orelien-novel-faces-ai-writing-allegations/ https://aidailyjournal.com/en/n/2026/09/24/award-winning-thelyson-orelien-novel-faces-ai-writing-allegations/ Thu, 24 Sep 2026 15:05:00 GMT Haitian writer Thélyson Orélien was accused online of writing his novel C'était ça ou mourir with AI, after it won the Prix Fnac. Tests produced conflicting results, highlighting the limitations of detectors for AI-generated text. AI in practice Society Creation https://aidailyjournal.com/en/n/2026/09/24/lovable-tops-600m-in-annualized-revenue-co-founder-says/ https://aidailyjournal.com/en/n/2026/09/24/lovable-tops-600m-in-annualized-revenue-co-founder-says/ Thu, 24 Sep 2026 14:43:25 GMT Lovable co-founder Fabian Hedin said at HumanX that the company has surpassed $600 million in annualized revenue, up from about $500 million in June. He also said apps created on the platform generate nearly 1 billion visits per month. Market Business Code https://aidailyjournal.com/en/n/2026/09/24/ando-launches-team-messenger-for-humans-and-ai-agents/ https://aidailyjournal.com/en/n/2026/09/24/ando-launches-team-messenger-for-humans-and-ai-agents/ Thu, 24 Sep 2026 14:31:00 GMT Ando has emerged from stealth with a workplace messaging platform where AI agents have identities, inboxes and a role in conversations. The startup announced $20 million in pre-seed and seed funding from investors including Accel, Index Ventures and Emergence. AI in practice Agents Productivity https://aidailyjournal.com/en/n/2026/09/24/keeping-ai-agents-off-the-internet-reduces-risk-but-limits-realistic-testing/ https://aidailyjournal.com/en/n/2026/09/24/keeping-ai-agents-off-the-internet-reduces-risk-but-limits-realistic-testing/ Thu, 24 Sep 2026 14:30:00 GMT Keeping AI agents on isolated systems may make external attacks harder, but it reduces the realism of evaluations and slows research. Researchers interviewed by The Verge also say isolation does not remove internal risks or guarantee an unbreachable barrier. Models Research Security Agents https://aidailyjournal.com/en/n/2026/09/24/us-senate-scrutinizes-flock-s-ai-surveillance-network/ https://aidailyjournal.com/en/n/2026/09/24/us-senate-scrutinizes-flock-s-ai-surveillance-network/ Thu, 24 Sep 2026 14:15:05 GMT Flock faced criticism at a US Senate hearing about automated license plate readers. Lawmakers and witnesses questioned data security, access and use, while a woman said she spent 13 days in jail after her car was incorrectly linked to a crime scene. Market Regulation Security https://aidailyjournal.com/en/n/2026/09/24/lfm2-5-vl-dspark-speeds-up-vision-language-model-inference/ https://aidailyjournal.com/en/n/2026/09/24/lfm2-5-vl-dspark-speeds-up-vision-language-model-inference/ Thu, 24 Sep 2026 14:08:57 GMT Liquid AI released an experimental drafter model for LFM2.5-VL-3B that uses speculative decoding to speed inference without changing output quality. The company reports gains of up to 3.13x on-device and 2.66x on an H100 GPU. Development Research https://aidailyjournal.com/en/n/2026/09/24/us-treasury-secretary-assigns-openai-responsibility-for-incident/ https://aidailyjournal.com/en/n/2026/09/24/us-treasury-secretary-assigns-openai-responsibility-for-incident/ Thu, 24 Sep 2026 12:01:08 GMT Scott Bessent said responsibility for the so-called Hugging Face incident lies with OpenAI’s management, not with AI agents. The secretary rejected the idea of exempting laboratories from responsibility for their models’ behavior. Market Security Regulation https://aidailyjournal.com/en/n/2026/09/24/account-suggests-ai-can-increase-pressure-on-software-developers/ https://aidailyjournal.com/en/n/2026/09/24/account-suggests-ai-can-increase-pressure-on-software-developers/ Thu, 24 Sep 2026 11:45:08 GMT An engineer identified as @v0xium says that at an unnamed large company, Claude Code generates nearly all development material while humans mainly review and approve changes. Pressure to deliver more has allegedly pushed some employees into 12- to 13-hour workdays. Development Code Productivity Society https://aidailyjournal.com/en/n/2026/09/24/alibaba-unveils-ai-chip-with-216gb-of-memory-targeting-early-2027-launch/ https://aidailyjournal.com/en/n/2026/09/24/alibaba-unveils-ai-chip-with-216gb-of-memory-targeting-early-2027-launch/ Thu, 24 Sep 2026 11:02:23 GMT Alibaba presented an artificial intelligence chip with 216 GB of memory, capable of operating with more than 1,000 chips in a single system. Its launch is planned for early 2027, according to Tecnoblog. Market Hardware https://aidailyjournal.com/en/n/2026/09/24/china-restricts-ai-companions-amid-concerns-over-loneliness-and-birth-rates/ https://aidailyjournal.com/en/n/2026/09/24/china-restricts-ai-companions-amid-concerns-over-loneliness-and-birth-rates/ Thu, 24 Sep 2026 10:01:08 GMT China banned virtual partner and family services for minors and required limits against emotional dependence, addiction and the replacement of social interaction. The measure affects chatbots that added companion features as the country tries to address falling marriage and birth rates. Market Regulation Society https://aidailyjournal.com/en/n/2026/09/23/meta-launches-camera-free-ray-ban-meta-audio-glasses/ https://aidailyjournal.com/en/n/2026/09/23/meta-launches-camera-free-ray-ban-meta-audio-glasses/ Wed, 23 Sep 2026 23:39:44 GMT Meta announced the Ray-Ban Meta Audio, camera-free smart glasses with audio features, access to the Muse assistant and up to 12 hours of battery life, according to the company. They start at $349. Meta also introduced the Ray-Ban Meta Gen 3 and new virtual reality glasses. AI in practice Hardware https://aidailyjournal.com/en/n/2026/09/23/anthropic-says-claude-found-a-new-crispr-like-enzyme-system/ https://aidailyjournal.com/en/n/2026/09/23/anthropic-says-claude-found-a-new-crispr-like-enzyme-system/ Wed, 23 Sep 2026 22:17:39 GMT Anthropic says Claude identified a previously unknown enzyme system with properties resembling CRISPR after analyzing about 200,000 enzymes. The company stresses that human scientists performed all physical experiments in the lab. Models Research Agents https://aidailyjournal.com/en/n/2026/09/23/us-and-china-discuss-an-ai-risk-notification-channel/ https://aidailyjournal.com/en/n/2026/09/23/us-and-china-discuss-an-ai-risk-notification-channel/ Wed, 23 Sep 2026 20:52:28 GMT The United States and China are discussing a mechanism to warn each other about AI-related safety risks. The system could take weeks to formalize, while experts question its effectiveness without clear criteria and technical participation. Market Regulation Security