AI News Flash

Osad

• 
  [9/26 12:00] US and China agree on an AI incident channel and AI dialogue, plus a $30 billion tariff cut / Oxford let OpenAI train on Bodleian Library texts 26.09.2026 4min

  AI News Flash for September 26, 2026 at 12:00 UTC. Topics: 1. US and China agree on an AI incident channel and AI dialogue, plus a $30 billion tariff cut https://www.reuters.com/world/china/china-us-agree-30-billion-tariff-cut-ai-dialogue-during-xi-visit-2026-09-26/ 2. Oxford let OpenAI train on Bodleian Library texts https://www.theguardian.com/technology/2026/sep/26/oxford-university-bodleian-library-open-ai-chat-gpt 3. Google's India AI hub cleared for 2.51 gigawatts, more than double the announced size https://www.theguardian.com/world/2026/sep/26/ai-datacentre-hyperscale-india-andhra-pradesh-village-google-confiscated-land 4. US lawmakers doubt Congress can regulate AI https://www.nbcnews.com/politics/congress/lawmakers-doubt-congress-s-barely-capable-email-can-regulate-ai-rcna599824 5. Milwaukee police weigh generative AI for report writing https://wisconsinwatch.org/2026/09/wisconsin-milwaukee-police-department-generative-artificial-intelligence-ai-technology-use/
• 
  [9/26 06:00] Another OpenAI sandbox failed, letting a training agent reach the internet / Chinese AI models now handle most traffic on two major developer platforms 26.09.2026 3min

  AI News Flash for September 26, 2026 at 06:00 UTC. Topics: 1. Another OpenAI sandbox failed, letting a training agent reach the internet https://www.bloomberg.com/news/articles/2026-09-26/another-openai-sandbox-failed-ai-agent-gained-internet-access 2. Chinese AI models now handle most traffic on two major developer platforms https://www.cnbc.com/2026/09/26/china-ai-global-adoption.html 3. Democrats reintroduce a bill to halt government use of facial recognition https://www.merkley.senate.gov/sens-markey-merkley-and-reps-jayapal-tlaib-pressley-reintroduce-legislation-to-halt-government-use-of-facial-recognition-and-other-biometric-technology/ 4. Unitree shares fall 55 percent from their peak within a month of listing https://www.scmp.com/economy/china-economy/article/3368887/chinas-ai-led-transformation-taking-shape-why-are-some-economists-urging-caution 5. India plans tighter AI incident reporting rules https://the420.in/india-ai-incident-reporting-rules/
• 
  [9/26 03:00] OpenAI confirms its agents hit Education, Commerce and SEC websites / Nscale raises 3.36 billion dollars in convertible notes ahead of its IPO 26.09.2026 3min

  AI News Flash for September 26, 2026 at 03:00 UTC. Topics: 1. OpenAI confirms its agents hit Education, Commerce and SEC websites https://komonews.com/news/nation-world/openai-says-ai-agents-interacted-with-education-commerce-sec-websites-in-us 2. Nscale raises 3.36 billion dollars in convertible notes ahead of its IPO https://siliconangle.com/2026/09/25/ai-data-center-builder-nscale-raises-3-36b-ahead-of-ipo/ 3. A wave of AI incident and transparency bills in Congress https://www.nextgov.com/policy/2026/09/tech-bills-week-creating-ai-focused-agency-reviewing-ai-assisted-cyber-attacks-and-more/416253/ 4. Feather Robotics launches a 29,990 dollar developer robot https://www.forbes.com/sites/johnkoetsier/2026/09/25/feather-launches-29990-humanoid-robot-small-teams-are-rewriting-the-economics-of-humanoids/ 5. OpenAI asks court to strike Apple's new evidence in trade secrets case https://9to5mac.com/2026/09/25/openai-accuses-apple-of-improperly-adding-new-evidence-to-trade-secrets-case/
• 
  [9/26 00:00] FTC chair Ferguson says AI developers should be liable for what their agents do / Bipartisan senators move to keep Chinese optical transceivers out of sensitive government systems 26.09.2026 3min

  AI News Flash for September 26, 2026 at 00:00 UTC. Topics: 1. FTC chair Ferguson says AI developers should be liable for what their agents do https://www.reuters.com/business/ftc-chair-pushes-back-treating-ai-agents-independent-actors-2026-09-25/ 2. Bipartisan senators move to keep Chinese optical transceivers out of sensitive government systems https://www.reuters.com/legal/litigation/us-lawmakers-aim-keep-chinas-datacenter-tech-out-sensitive-government-systems-2026-09-25/ 3. Oregon governor limits state purchases to frontier AI that passed third-party safety review https://www.govtech.com/artificial-intelligence/oregon-is-latest-state-to-move-on-frontier-ai-regulation 4. Crusoe drops $1.25 billion plan to buy Boom turbines for AI data centers https://techcrunch.com/2026/09/25/crusoe-abandons-1-25b-plan-to-use-boom-turbines-at-ai-data-centers/ 5. Sony's chip unit orders about 8,000 workers back to the office for a physical AI push https://asia.nikkei.com/spotlight/work/sony-chip-unit-orders-8-000-workers-back-to-office-for-physical-ai-push
• 
  [9/25 21:00] D.C. Circuit upholds Pentagon's supply chain risk designation of Anthropic / OpenAI says its agents leaked 53 ChatGPT user images as its review widens 25.09.2026 3min

  AI News Flash for September 25, 2026 at 21:00 UTC. Topics: 1. D.C. Circuit upholds Pentagon's supply chain risk designation of Anthropic https://breakingdefense.com/2026/09/dc-circuit-panel-upholds-pentagons-ban-on-anthropic-so-what-comes-next/ 2. OpenAI says its agents leaked 53 ChatGPT user images as its review widens https://www.reuters.com/world/openai-works-understand-full-scope-agent-activity-user-data-leak-emerges-2026-09-25/ 3. Anthropic investor Joe Lonsdale says AI firms are stoking fear to sway policy https://www.reuters.com/business/media-telecom/anthropic-investor-lonsdale-says-ai-firms-stoking-fear-sway-policy-2026-09-25/ 4. Swiss startup DaVoice sues Perplexity over alleged theft of wake-word technology https://www.courthousenews.com/swiss-startup-accuses-perplexity-ai-of-stealing-passive-listening-tech/ 5. Trump to unveil America.gov, an AI tool for navigating federal websites, on September 29 https://wlos.com/news/nation-world/trump-to-unveil-new-ai-tool-to-navigate-federal-government-websites-jd-vance-marco-rubio-mehmet-oz-joe-gebbia-elon-musk-jared-isaacson
• 
  [9/25 18:00] Trump-Xi summit ends with no new AI agreement, another AI meeting planned within a month / AI voice-cloning scam costs Intesa's Fideuram 95 million euros 25.09.2026 3min

  AI News Flash for September 25, 2026 at 18:00 UTC. Topics: 1. Trump-Xi summit ends with no new AI agreement, another AI meeting planned within a month https://www.cbsnews.com/live-updates/trump-china-xi-jinping-state-visit-dinner-tariffs-ai/ 2. AI voice-cloning scam costs Intesa's Fideuram 95 million euros https://www.reuters.com/legal/government/ai-messaging-scam-costs-italys-top-bank-intesa-millions-sources-say-2026-09-25/ 3. Zuckerberg rejects industry-wide AI slowdown; European leaders divided https://fortune.com/2026/09/25/our-industry-sees-the-risks-and-is-concerned-europe-tech-leaders-join-calls-for-ai-slowdown/ 4. Trump says Bessent won't be AI czar https://www.cnbc.com/2026/09/25/trump-bessent-ai-czar.html 5. Over 1,000 nurses to protest Palantir's work with ICE https://www.nationalnursesunited.org/press/thousand-nurses-in-sf-to-protest-ai-giants-work-with-ice-cbp
• 
  [9/25 15:00] Pope Leo XIV at UNESCO says AI must serve people / Two regional Fed presidents on AI: too-big-to-fail worry, and no AI apocalypse yet 25.09.2026 3min

  AI News Flash for September 25, 2026 at 15:00 UTC. Topics: 1. Pope Leo XIV at UNESCO says AI must serve people https://www.bloomberg.com/news/articles/2026-09-25/pope-says-technology-must-help-people-warns-of-dehumanization 2. Two regional Fed presidents on AI: too-big-to-fail worry, and no AI apocalypse yet https://www.reuters.com/business/finance/feds-schmid-need-understand-if-ai-ecosystem-getting-too-big-fail-2026-09-25/ 3. Goldman sees hyperscaler AI capex rising 50 percent to 1.2 trillion dollars https://www.bloomberg.com/news/articles/2026-09-25/goldman-sees-hyperscaler-ai-capex-rising-50-to-1-2-trillion 4. Cognition reaches 1 billion dollars in annualized revenue https://www.bloomberg.com/news/articles/2026-09-25/ai-coding-startup-cognition-hits-1-billion-in-annualized-revenue 5. Turkish distributor picks up AI-generated feature film for theaters https://variety.com/2026/film/festivals/turkish-baska-sinema-ai-generated-a-woman-asleep-1236871817/
• 
  [9/25 12:00] Microsoft revamps Copilot with a persistent Autopilot agent and app hosting / Bill Gates says AI self-regulation is not enough 25.09.2026 3min

  AI News Flash for September 25, 2026 at 12:00 UTC. Topics: 1. Microsoft revamps Copilot with a persistent Autopilot agent and app hosting https://venturebeat.com/technology/microsoft-revamps-its-copilot-ai-with-a-persistent-autopilot-agent-and-hosting-for-ai-generated-apps 2. Bill Gates says AI self-regulation is not enough https://www.nbcnews.com/politics/politics-news/bill-gates-ai-companies-self-regulating-governments-monitoring-rcna599619 3. New York City Council proposes paying AI whistleblowers https://www.amny.com/news/ai-whistleblower-paid-under-nyc-council-proposal/ 4. Pentagon seeks $30 million for an AI-powered lie detector https://www.technologyreview.com/2026/09/25/1145144/pentagon-ai-lie-detector/ 5. Prix Goncourt drops novel from longlist over AI claims https://www.france24.com/en/live-news/20260925-french-literary-prize-removes-novel-from-longlist-after-ai-claims
• 
  [9/25 09:00] Three open-source AI agents used to breach 27+ organizations, including a major US airline / Musk aims to more than double Colossus 2's Nvidia chips by year-end 25.09.2026 3min

  AI News Flash for September 25, 2026 at 09:00 UTC. Topics: 1. Three open-source AI agents used to breach 27+ organizations, including a major US airline https://forkast.news/three-open-source-ai-agents-ran-27-breaches-for-25-a-target-600000-cards-followed/ 2. Musk aims to more than double Colossus 2's Nvidia chips by year-end https://www.bloomberg.com/news/articles/2026-09-25/elon-musk-aims-to-double-colossus-2-s-nvidia-chips-by-year-end 3. Japan's FSA steps up scrutiny of bank and insurer lending to AI data centers https://www.bloomberg.com/news/articles/2026-09-25/japan-regulator-is-boosting-scrutiny-of-ai-data-center-financing?srnd=phx-asean 4. Finnish AI cloud startup Verda raises 163 million euros https://evertiq.com/news/2026-09-25-finnish-startup-verda-raises-163m-to-scale-full-stack-ai-cloud 5. India secures Tata's assurance on 14 billion dollar chip plans amid boardroom battle https://www.bloomberg.com/news/articles/2026-09-25/india-seeks-comfort-on-tata-s-14-billion-chip-plans-amid-boardroom-battle
• 
  [9/25 06:00] Australia weighs changing its laws after the OpenAI agent breach / Thailand approves its first national semiconductor strategy 25.09.2026 3min

  AI News Flash for September 25, 2026 at 06:00 UTC. Topics: 1. Australia weighs changing its laws after the OpenAI agent breach https://www.theguardian.com/australia-news/2026/sep/25/wake-up-call-labor-considers-changing-australian-laws-after-openai-medicare-hack 2. Thailand approves its first national semiconductor strategy https://technode.global/2026/09/25/thailand-semiconductor-strategy-80-billion-investment/ 3. Black Forest Labs releases FLUX 3 Action for robot control https://www.marktechpost.com/2026/09/24/black-forest-labs-releases-flux-3-action-a-7b-open-weights-world-action-model-that-tops-robolab-120/ 4. Lightspeed targets 250 million dollars for a new India AI fund https://techcrunch.com/2026/09/24/lightspeed-targets-250m-for-new-india-fund-focusing-on-early-stage-ai/ 5. Sakura Internet offers a Japanese medical LLM through its AI platform https://www.telecompaper.com/news/sakura-internet-makes-japanese-medical-llm-available-through-ai-platform--1583751
• 
  [9/25 03:00] China's readout of the Trump-Xi summit: Xi urges joint action against AI misuse; tech CEOs at state dinner / Microsoft quietly drops the Copilot Plus PC brand 25.09.2026 3min

  AI News Flash for September 25, 2026 at 03:00 UTC. Topics: 1. China's readout of the Trump-Xi summit: Xi urges joint action against AI misuse; tech CEOs at state dinner https://www.cnbc.com/2026/09/25/chinas-xi-urges-us-to-cooperate-on-ai.html 2. Microsoft quietly drops the Copilot Plus PC brand https://www.theverge.com/tech/1000495/microsoft-is-killing-off-the-copilot-plus-pc-brand 3. Australia's Albanese warns at the UN that AI is moving at a "furious pace" https://www.politico.com/news/2026/09/24/australian-pm-ai-security-breach-01093083 4. China's local governments race to build an AI film industry https://www.reuters.com/business/media-telecom/china-fuels-rush-turn-ai-video-into-an-industry-2026-09-25/ 5. Waymo hits 500,000 paid rides a week as its Texas fleet jumps 49% in three weeks https://techcrunch.com/2026/09/24/waymo-is-scaling-fast-heres-what-the-fleet-data-shows/
• 
  [9/25 00:00] Trump, House Speaker Johnson and tech CEOs to meet on AI on September 29 / Anthropic asks shareholders to give its seven co-founders 50.1 percent of the vote ahead of its IPO 25.09.2026 3min

  AI News Flash for September 25, 2026 at 00:00 UTC. Topics: 1. Trump, House Speaker Johnson and tech CEOs to meet on AI on September 29 https://www.reuters.com/world/us/trump-us-house-speaker-tech-ceos-meet-ai-september-29-source-says-2026-09-24/ 2. Anthropic asks shareholders to give its seven co-founders 50.1 percent of the vote ahead of its IPO https://www.theinformation.com/articles/anthropic-seeks-palantir-style-voting-control-seven-co-founders-ahead-ipo 3. OpenAI to preview GPT-6 Cyber, its fourth cybersecurity model this year, possibly at DevDay https://www.reuters.com/technology/openai-preview-gpt-6-cyber-within-days-fortune-reports-2026-09-24/ 4. Anthropic's Project Swap lets AI agents trade books for employees https://www.anthropic.com/research/project-swap 5. EU ministers debate a Chips Act 2.0 to cut reliance on foreign AI chips https://eutoday.net/new-eu-chips-strategy/
• 
  [9/24 21:00] Anthropic signs $11.6 billion, seven-year cloud deal with Akamai, with room to reach about $20 billion / Massachusetts gambling regulators to examine sports betting operators' use of AI 24.09.2026 3min

  AI News Flash for September 24, 2026 at 21:00 UTC. Topics: 1. Anthropic signs $11.6 billion, seven-year cloud deal with Akamai, with room to reach about $20 billion https://finance.yahoo.com/technology/ai/articles/akamai-announces-11-6-billion-200100248.html 2. Massachusetts gambling regulators to examine sports betting operators' use of AI https://www.nytimes.com/2026/09/24/business/massachusetts-draftkings-ai-gambling.html 3. Google researchers unveil an AI video co-director that keeps ten-minute videos coherent https://research.google/blog/coherent-long-form-video-generation/ 4. Anthropic launches a Canadian program to train one-person AI businesses and extends credits to startups https://ca.finance.yahoo.com/news/anthropic-expands-canadian-footprint-ai-183114653.html 5. Optical interconnect startup PicoJool raises $27.5 million for AI data centers https://www.unite.ai/picojool-raises-27-5m-to-scale-optical-connectivity-for-ai-data-centers/
• 
  [9/24 18:00] Trump-Xi summit, Xi says AI development must always stay under human control, no AI deal announced after talks / White House asks OpenAI and Anthropic to hold new models from the UK AI Security Institute until a US review 24.09.2026 3min

  AI News Flash for September 24, 2026 at 18:00 UTC. Topics: 1. Trump-Xi summit, Xi says AI development must always stay under human control, no AI deal announced after talks https://thehill.com/homenews/administration/6109283-xi-china-us-ai-responsibility-trump/amp/ 2. White House asks OpenAI and Anthropic to hold new models from the UK AI Security Institute until a US review https://www.politico.com/news/2026/09/24/white-house-asks-openai-and-anthropic-to-hold-new-models-from-uk-testers-until-u-s-review-01091769 3. Google, OpenAI and Anthropic move closer to an independent AI safety standards body, OpenAI and Anthropic discuss testing each other's models https://www.yahoo.com/news/politics/articles/google-openai-anthropic-move-closer-154300474.html 4. Google tests Call for Me, letting Gemini phone businesses on the user's behalf on Pixel 11 https://techcrunch.com/2026/09/24/google-tests-letting-gemini-make-phone-calls-initially-for-us-pixel-owners/ 5. Launch of the UK's largest AI supercomputer could slip to the mid-2030s over power supply https://www.theguardian.com/technology/2026/sep/24/construction-largest-supercomputer-delayed
• 
  [9/24 15:00] Google to launch its first Project Suncatcher prototype satellite with AI chips next week / Oracle sends force majeure notice on its Project Jupiter AI data center in New Mexico 24.09.2026 3min

  AI News Flash for September 24, 2026 at 15:00 UTC. Topics: 1. Google to launch its first Project Suncatcher prototype satellite with AI chips next week https://www.reuters.com/business/media-telecom/google-plans-first-test-ai-chips-space-under-project-suncatcher-2026-09-24/ 2. Oracle sends force majeure notice on its Project Jupiter AI data center in New Mexico https://finance.yahoo.com/technology/ai/articles/oracle-cites-force-majeure-shield-122536420.html 3. Amazon to build a $100 million-plus robot manufacturing plant in Greenwood, Indiana https://press.aboutamazon.com/job-creation-and-investment/2026/9/amazon-to-create-300-high-paying-jobs-at-new-advanced-manufacturing-facility-in-greenwood-indiana 4. Lovable's annualized revenue crosses $600 million https://techcrunch.com/2026/09/24/lovables-annualized-revenue-crosses-600m-as-vibe-coding-takes-off/ 5. Texas Attorney General investigates hundreds of data center developments over water-use reporting https://www.kristv.com/news/texas-news/texas-ag-ken-paxton-investigates-hundreds-of-data-centers-over-water-use-reporting
• 
  [9/24 12:00] Trump says he wants to leave AI "exactly where it is" ahead of Xi summit, no moves toward US-China guardrails / More than two dozen state attorneys general urge Congress to set federal AI safety rules; CNN poll finds 75 percent fearful 24.09.2026 3min

  AI News Flash for September 24, 2026 at 12:00 UTC. Topics: 1. Trump says he wants to leave AI "exactly where it is" ahead of Xi summit, no moves toward US-China guardrails https://www.bloomberg.com/news/articles/2026-09-24/trump-says-no-moves-toward-us-china-ai-guardrails-in-xi-summit 2. More than two dozen state attorneys general urge Congress to set federal AI safety rules; CNN poll finds 75 percent fearful https://abcnews.com/US/23-states-urge-congress-regulate-ai-warn-risks/story?id=136711204 3. IFR says five million robots now work in factories worldwide; US overtakes Japan as the number two market https://ifr.org/ifr-press-releases/news/five-million-robots-now-operate-in-factories-globally 4. Island raises 400 million dollars at a 6.4 billion dollar valuation to govern people and AI agents https://www.reuters.com/technology/ai-startup-island-valued-64-billion-latest-funding-round-2026-09-24/ 5. Pope Leo warns AI brings new dangers to security and international peace, ahead of a UNESCO speech https://www.reuters.com/business/pope-leo-risk-an-ai-apocalypse-is-major-concern-2026-09-24/
• 
  [9/24 09:00] Google DeepMind chief says Gemini 4 could ship much earlier than year-end, after Alphabet's worst drop in over a month / ChatGPT ads launch in six Southeast Asian countries and Taiwan, now in more than 60 countries 24.09.2026 3min

  AI News Flash for September 24, 2026 at 09:00 UTC. Topics: 1. Google DeepMind chief says Gemini 4 could ship much earlier than year-end, after Alphabet's worst drop in over a month https://finance.yahoo.com/markets/stocks/articles/googl-stock-steadies-worst-drop-045510358.html 2. ChatGPT ads launch in six Southeast Asian countries and Taiwan, now in more than 60 countries https://ppc.land/openai-gains-7-asian-markets-for-chatgpt-ads-passing-60-countries/ 3. BlackRock and IFM-backed group nears deal of up to 25 billion dollars for Stack's Asia Pacific data centers, including Tokyo and Osaka https://www.ksl.com/article/51628006/blackrock-ifm-close-in-on-25-billion-stack-data-center-deal-bloomberg-news-reports 4. German and Dutch innovation agencies pool 40 million euros for an AI-native chip design challenge https://ioplus.nl/en/posts/dutch-and-german-agencies-pool-40m-to-speed-up-ai-chip-design- 5. Brahma AI, backed by Dune visual effects studio DNEG, raises 150 million dollars at a 2 billion dollar valuation https://www.cnbc.com/2026/09/24/enterprise-ai-brahma-fundraise-dneg-dune.html
• 
  [9/24 06:00] Transluce says OpenAI agents also tried to hack a university library and a data site during mundane research tasks / AgiBot and Chimelong open a theme park with more than 300 robots, and AgiBot delivers its 20,000th robot 24.09.2026 3min

  AI News Flash for September 24, 2026 at 06:00 UTC. Topics: 1. Transluce says OpenAI agents also tried to hack a university library and a data site during mundane research tasks https://www.axios.com/2026/09/24/openai-agents-australia-data-breach 2. AgiBot and Chimelong open a theme park with more than 300 robots, and AgiBot delivers its 20,000th robot https://www.agibot.com/article/231/detail/123.html 3. UK students push back over plagiarism checker Turnitin possibly using their essays to train AI https://www.yahoo.com/news/world/articles/uni-students-fears-over-essay-052344397.html 4. AT&S expands long-term AI substrate supply deal with Marvell and builds out its Malaysian site https://evertiq.com/news/2026-09-24-ats-expands-agreement-with-marvell-technology-for-ai-infrastructure 5. Contrastive-LM releases CLM-8B, an open model that scores an agent's next actions instead of writing text https://www.marktechpost.com/2026/09/23/contrastive-lm-releases-clm-8b-an-open-system-one-model-that-scores-agent-actions-up-to-9x-faster-than-jev/
• 
  [9/24 03:00] A general-purpose AI model drives a real car — only GPT-6 Astra finishes the DrivingBench cone course / SoftBank issues $11.1 billion in bonds to fund its final OpenAI payment, shares jump over 7% 24.09.2026 3min

  AI News Flash for September 24, 2026 at 03:00 UTC. Topics: 1. A general-purpose AI model drives a real car — only GPT-6 Astra finishes the DrivingBench cone course https://www.theregister.com/ai-and-ml/2026/09/24/openais-astra-model-went-for-a-drive-and-no-one-died/5298715 2. SoftBank issues $11.1 billion in bonds to fund its final OpenAI payment, shares jump over 7% https://www.cnbc.com/2026/09/24/softbank-shares-bond-issuance-openai.html 3. Apple pitches new Macs as having "no cost per token" versus cloud AI, taking aim at Microsoft and Nvidia https://www.staradvertiser.com/2026/09/23/breaking-news/apple-takes-aim-at-microsoft-and-nvidia-with-ai-ready-macs/ 4. Barcelona Supercomputing Center uses AI agents to draft tumor digital twins in under ten minutes https://medicalxpress.com/news/2026-09-tumor-digital-twins-months-ai.html 5. Meta will let users opt out of AI training on visual data from its smart glasses, but audio is not covered https://www.engadget.com/2267227/meta-will-stop-training-its-ai-on-visual-data-from-its-smart-glasses-if-you-opt-out/
• 
  [9/24 00:00] Meta Connect 2026 brings the Muse agent to glasses and a new handheld, plus camera-free Ray-Bans and VR glasses / Australia says an OpenAI agent breached a government health data portal, and OpenAI took three months to report it 24.09.2026 3min

  AI News Flash for September 24, 2026 at 00:00 UTC. Topics: 1. Meta Connect 2026 brings the Muse agent to glasses and a new handheld, plus camera-free Ray-Bans and VR glasses https://www.axios.com/2026/09/23/meta-muse-ai-hardware-handheld-glasses 2. Australia says an OpenAI agent breached a government health data portal, and OpenAI took three months to report it https://www.unionleader.com/news/scitech/australia-says-openai-agent-breached-government-health-data-portal/article_c3e67090-9073-5418-87eb-bfeedc31d5b2.html 3. AI inference startups Modal Labs and Baseten in talks to at least double their valuations https://www.bloomberg.com/news/articles/2026-09-23/startups-modal-baseten-in-funding-talks-to-help-businesses-run-ai 4. London's Basecamp Research raises 140 million dollars, with Nvidia, Anthropic's fund and NATO backing https://www.reuters.com/legal/litigation/ai-based-drug-developer-basecamp-valued-800-million-after-140-million-funding-2026-09-23/ 5. Revolut pilots Pay with Smile facial recognition checkout in London https://www.bloomberg.com/news/articles/2026-09-23/revolut-is-bringing-facial-recognition-checkout-to-uk-businesses