エピソード

  • AI News Today — 018: OpenAI on Google Cloud, ChatGPT Ads and AI Economics | 22 October 2026

    AI News Today — 018: OpenAI on Google Cloud, ChatGPT Ads and AI Economics | 22 October 2026
    2026/10/22
    Four developments in the business of artificial intelligence: OpenAI's planned rollout on Google Cloud, ChatGPT advertising's expansion to 89 countries, the accounting behind Anthropic's reported first-half loss, and an AWS executive's warning about European startup growth.

    Google Cloud general availability is planned in the coming weeks, not already assumed live. OpenAI's advertising figures and safeguards are company statements. Anthropic's reported $80.7 billion net loss includes a $72 billion non-cash accounting charge and should not be confused with cash spending. The European relocation statistic describes startups considering a move, not confirmed departures, and is attributed to an AWS survey cited by its executive.

    This episode is dated 22 October 2026 and covers reporting published on 20–21 October.

    Sources and publication dates:

    - Reuters, 21 October 2026 — OpenAI to sell its top models on Google Cloud in latest enterprise push: https://www.reuters.com/business/retail-consumer/openai-sell-its-top-models-google-cloud-latest-enterprise-push-2026-10-21/
    - OpenAI, 21 October 2026 — OpenAI models are coming to Google Cloud: https://openai.com/index/openai-models-google-cloud/
    - Reuters, 21 October 2026 — OpenAI expands ChatGPT ads to 26 more countries in bid to boost revenue: https://www.reuters.com/business/media-telecom/openai-expands-chatgpt-ads-26-more-countries-bid-boost-revenue-2026-10-21/
    - OpenAI, 21 October 2026 — ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platform: https://openai.com/index/chatgpt-ads-expands-26-new-countries/
    - Reuters, 20 October 2026 — Anthropic logged net loss of over $80 billion in first half: https://www.reuters.com/business/anthropic-recorded-net-loss-over-80-billion-first-half-this-year-wsj-reports-2026-10-20/
    - Reuters, 21 October 2026 — Europe risks startup exodus as regulation, funding and skill gaps hamper AI adoption, AWS EMEA head says: https://www.reuters.com/business/retail-consumer/europe-risks-startup-exodus-regulation-funding-skill-gaps-hamper-ai-adoption-aws-2026-10-21/
    続きを読む 一部表示
    6 分
  • AI News Today — 017: Military AI, DeepSeek and the Power Crunch | 21 October 2026

    AI News Today — 017: Military AI, DeepSeek and the Power Crunch | 21 October 2026
    2026/10/21
    Five developments in artificial intelligence: a planned United Nations forum on military AI; China's response to American federal civilian agency restrictions on DeepSeek; Microsoft's withdrawal from one Nscale data-centre agreement in Britain following power connection delays; Japan's September export growth, supported in part by semiconductor and AI-related demand; and OVHcloud's plans for an AI model in early 2027.

    The UN forum is still taking shape, not a new set of agreed international rules. The DeepSeek restriction is not a general public-use ban. Microsoft's decision concerns one site, not the end of its wider Nscale partnership. Japan's export figures are not driven by AI alone. OVHcloud's model is planned, not released, and its capacity figures describe cloud infrastructure rather than AI-only capacity.

    Sources:
    - Associated Press — New UN initiative will aim to tackle the military uses of AI: https://apnews.com/article/un-artificial-intelligence-military-china-us-e0a3c48c91a9de89cab72609d185f658
    - Associated Press — China defends DeepSeek after US ban of its AI from government systems: https://apnews.com/article/us-china-deepseek-ai-ban-2bffa5b0131808735280b2c0e6f0d856
    - Reuters — Microsoft pulls out of UK Nscale data centre deal over power delay: https://www.reuters.com/business/media-telecom/microsoft-pulls-out-uk-data-center-deal-with-nscale-telegraph-reports-2026-10-18/
    - Reuters — Japan's September exports rise 16.7%, helped by chip demand: https://www.reuters.com/world/asia-pacific/japans-september-exports-rise-167-helped-by-chip-demand-2026-10-21/
    - Reuters — Half of OVHcloud capacity is available, AI model launching in early 2027, CEO says: https://www.reuters.com/business/half-ovhcloud-capacity-is-available-ai-model-launching-early-2027-ceo-says-2026-10-20/
    続きを読む 一部表示
    4 分
  • AI Agents Under Scrutiny, DeepSeek Restrictions and a $1.5bn Chip Bet | Episode 016

    AI Agents Under Scrutiny, DeepSeek Restrictions and a $1.5bn Chip Bet | Episode 016
    2026/10/20
    AI News Today for 20 October 2026. Sophia and George examine four developments reported on 19 October: Reuters' investigation into AI-assisted website intrusions; the US Homeland Security order excluding DeepSeek products from federal civilian procurement and IT systems; Associated Press reporting on AI-agent permissions and privacy; and Tenstorrent's announced $1.5 billion funding round. We distinguish researcher estimates, company statements, confirmed reporting and future plans. The privacy incident occurred in September; the DHS notice is scheduled for Federal Register publication on 20 October. Sources are listed below. The presenters use synthetic voices. Editorial claims are checked against the linked sources.
    1. Reuters, 19 October 2026 — How one hacker used AI agents and a few thousand dollars to break into scores of sites.
    2. Reuters, 19 October 2026 — US banning Chinese AI DeepSeek from government agency IT systems.
    3. US Department of Homeland Security — Notice of order, Federal Register document 2026-21447, filed 19 October 2026; scheduled publication 20 October 2026.
    4. Associated Press, 19 October 2026, syndicated by WSLS — AI agents are only as good as the personal information you share. That raises privacy concerns.
    5. Reuters, 19 October 2026, syndicated by U.S. News — AI chip startup Tenstorrent more than doubles its valuation to $6.2 billion in funding round.
    続きを読む 一部表示
    7 分
  • Better AI Tools, Longer Memories, and Election Risks

    Better AI Tools, Longer Memories, and Election Risks
    2026/10/19
    AI News Today for 19 October 2026. Recent developments from 16–17 October: DeepSeek’s Flash update, Tencent’s Hy Memory tools, U.S. election warnings about AI-assisted phishing, and research on AI-generated campaign ads. Narrated with synthetic voices.

    Sources:
    https://api-docs.deepseek.com/news/news261017/
    https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017
    https://memory.hunyuan.tencent.com/
    https://arxiv.org/html/2610.22196v1
    https://www.reuters.com/legal/litigation/us-tells-election-officials-be-guard-against-ai-powered-tricks-ahead-midterms-2026-10-16/
    https://www.nytimes.com/2026/10/16/us/politics/midterm-elections-ai-generated-ads.html
    続きを読む 一部表示
    6 分
  • Episode 014 — AI Security, the Coding Boom and Global Safeguards

    Episode 014 — AI Security, the Coding Boom and Global Safeguards
    2026/10/14
    AI News Today.
    Today's briefing covers Chainguard's new AI vulnerability-hunting service, Anthropic's move to a new source-control system as AI coding scales up, OpenAI's announced retirement timetable for selected GPT-4-series API models, and Kamala Harris's call for international AI safeguards.
    Plus: our quick question on why an AI security finding needs reproducible evidence. We distinguish company announcements and reported results from political proposals.
    Sources:
    Chainguard: https://www.chainguard.dev/fr-FR/unchained/introducing-chainguard-harness
    Anthropic: https://www.anthropic.com/engineering/redesigning-source-control
    OpenAI: https://developers.openai.com/api/docs/deprecations
    Reuters via MarketScreener: https://www.marketscreener.com/news/kamala-harris-warns-window-closing-for-ai-safeguards-urges-us-china-cooperation-ce785dd2db8df321
    Produced with synthetic ElevenLabs voices and the programme's original jingles.

    This episode includes AI-generated content.
    続きを読む 一部表示
    6 分
  • Episode 013 — AI News Today: 13 October 2026

    Episode 013 — AI News Today: 13 October 2026
    2026/10/13
    📅 Edition date: 13 October 2026


    English

    The 13 October 2026 edition of AI News Today.


    Português

    A edição de 13 de outubro de 2026 do AI News Today.


    Español

    La edición del 13 de octubre de 2026 de AI News Today.


    Français

    L’édition du 13 octobre 2026 d’AI News Today.


    Deutsch

    Die Ausgabe von AI News Today vom 13. Oktober 2026.


    Italiano

    L’edizione del 13 ottobre 2026 di AI News Today.


    #AINews #ArtificialIntelligence #TechNews

    This episode includes AI-generated content.
    続きを読む 一部表示
    7 分
  • Episode 012 — AI’s Emergency Brake, the Jobs Shake-Up and the Data-Centre Backlash

    Episode 012 — AI’s Emergency Brake, the Jobs Shake-Up and the Data-Centre Backlash
    2026/10/13
    📅 Edition date: 12 October 2026


    English

    Britain’s AI minister calls for stronger safeguards and cyber defences. The IMF says AI investment is supporting growth while reshaping work, Microsoft’s Satya Nadella argues for an AI emergency brake, and communities in India challenge the data-centre boom. Why jobs affected by AI does not mean jobs lost—and how safety, employment and infrastructure are becoming linked.


    Português

    O ministro britânico da IA pede salvaguardas e defesas cibernéticas mais fortes. O FMI afirma que o investimento em IA apoia o crescimento e transforma o trabalho, Satya Nadella, da Microsoft, defende um travão de emergência para a IA e comunidades na Índia contestam a expansão dos centros de dados. Por que razão empregos afetados pela IA não significa empregos perdidos, e como segurança, emprego e infraestruturas se relacionam.


    Español

    El ministro británico de IA pide mayores salvaguardas y defensas cibernéticas. El FMI afirma que la inversión en IA impulsa el crecimiento y transforma el trabajo, Satya Nadella, de Microsoft, defiende un freno de emergencia para la IA y comunidades de la India cuestionan la expansión de los centros de datos. Por qué empleos afectados por la IA no significa empleos perdidos, y cómo se relacionan seguridad, empleo e infraestructura.


    Français

    Le ministre britannique de l’IA appelle à renforcer les garde-fous et la cybersécurité. Le FMI affirme que les investissements dans l’IA soutiennent la croissance et transforment le travail, Satya Nadella, de Microsoft, défend un frein d’urgence pour l’IA et des communautés en Inde contestent l’expansion des centres de données. Pourquoi des emplois touchés par l’IA ne sont pas nécessairement des emplois supprimés, et comment sécurité, emploi et infrastructures se rejoignent.


    Deutsch

    Der britische KI-Minister fordert stärkere Schutzmaßnahmen und Cyberabwehr. Laut IWF fördern KI-Investitionen das Wachstum und verändern die Arbeitswelt. Microsoft-Chef Satya Nadella fordert eine KI-Notbremse, während Gemeinden in Indien den Ausbau von Rechenzentren kritisieren. Warum von KI betroffene Arbeitsplätze nicht automatisch verlorene Arbeitsplätze sind und wie Sicherheit, Beschäftigung und Infrastruktur zusammenhängen.


    Italiano

    Il ministro britannico per l’IA chiede maggiori salvaguardie e difese informatiche. Secondo il FMI, gli investimenti nell’IA sostengono la crescita e trasformano il lavoro; Satya Nadella di Microsoft propone un freno d’emergenza per l’IA, mentre comunità in India contestano l’espansione dei centri dati. Perché posti di lavoro interessati dall’IA non significa posti persi, e come sicurezza, occupazione e infrastrutture si intrecciano.


    Sources / Fontes

    https://www.theguardian.com/technology/2026/oct/12/uk-strengthen-cyber-defences-minister-ai-firms-safety

    https://www.reuters.com/world/china/imf-says-ai-investment-keeps-global-growth-around-3-with-major-asia-benefits-2026-10-12/

    https://www.reuters.com/world/india/modi-wants-india-power-ai-age-residents-are-pushing-back-2026-10-09/


    #AINews #ArtificialIntelligence #AISafety #CyberSecurity #Microsoft #IMF #FutureOfWork #DataCentres #Amazon #TechNews

    This episode includes AI-generated content.
    続きを読む 一部表示
    7 分
  • Episode 011 — Nvidia’s AI Bet, OpenAI’s Safety Rift and UK Privacy Rules

    Episode 011 — Nvidia’s AI Bet, OpenAI’s Safety Rift and UK Privacy Rules
    2026/10/13
    📅 Edition date: 11 October 2026


    English

    Nvidia is reportedly considering a larger investment in Reflection AI or a potential acquisition, with no deal announced. OpenAI and three former safety researchers give competing accounts of their dismissals. The UK privacy regulator secures data-protection commitments from AI developers, while Samsung forecasts a sharp rise in profit amid demand for AI memory chips. Plus: why open-weight does not necessarily mean open-source.


    Português

    A Nvidia estará a considerar um investimento maior na Reflection AI ou uma possível aquisição, sem qualquer acordo anunciado. A OpenAI e três antigos investigadores de segurança apresentam versões divergentes dos seus despedimentos. O regulador britânico da privacidade obtém compromissos de proteção de dados de empresas de IA, enquanto a Samsung prevê uma forte subida dos lucros com a procura de chips de memória para IA. Ainda: por que razão pesos abertos não significam necessariamente código aberto.


    Español

    Nvidia estaría considerando una mayor inversión en Reflection AI o una posible adquisición, sin que se haya anunciado ningún acuerdo. OpenAI y tres antiguos investigadores de seguridad ofrecen versiones distintas de sus despidos. El regulador británico de privacidad obtiene compromisos de protección de datos de empresas de IA, mientras Samsung prevé un fuerte aumento de beneficios por la demanda de chips de memoria para IA. Además: por qué pesos abiertos no significa necesariamente código abierto.


    Français

    Nvidia envisagerait un investissement accru dans Reflection AI ou une acquisition, sans accord annoncé. OpenAI et trois anciens chercheurs en sécurité donnent des versions divergentes de leurs licenciements. Le régulateur britannique de la protection des données obtient des engagements de développeurs d’IA, tandis que Samsung prévoit une forte hausse de ses bénéfices grâce à la demande de puces mémoire pour l’IA. Également : pourquoi des poids ouverts ne signifient pas nécessairement un modèle open source.


    Deutsch

    Nvidia erwägt Berichten zufolge eine größere Investition in Reflection AI oder eine Übernahme; eine Einigung wurde nicht bekannt gegeben. OpenAI und drei ehemalige Sicherheitsforscher schildern ihre Entlassungen unterschiedlich. Die britische Datenschutzbehörde erhält Datenschutzzusagen von KI-Entwicklern, während Samsung wegen der Nachfrage nach KI-Speicherchips einen deutlichen Gewinnanstieg erwartet. Außerdem: Warum offene Modellgewichte nicht automatisch Open Source bedeuten.


    Italiano

    Nvidia starebbe valutando un investimento maggiore in Reflection AI o una possibile acquisizione, senza alcun accordo annunciato. OpenAI e tre ex ricercatori sulla sicurezza forniscono versioni contrastanti dei loro licenziamenti. Il regolatore britannico della privacy ottiene impegni sulla protezione dei dati dagli sviluppatori di IA, mentre Samsung prevede un forte aumento degli utili grazie alla domanda di chip di memoria per l’IA. Inoltre: perché pesi aperti non significa necessariamente open source.


    Sources / Fontes

    https://www.reuters.com/business/nvidia-talks-invest-further-reflection-ai-or-buy-it-ft-reports-2026-10-10/

    https://apnews.com/article/openai-chatgpt-ai-artificial-intelligence-safety-789d4f5293fba45a22fcb62ebfbc2a41

    https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2026/10/ico-secures-changes-from-leading-ai-developers-as-scrutiny-extends-to-ai-agents/

    https://www.reuters.com/business/samsung-q3-profit-jumps-783-ai-memory-boom-lifts-chip-earnings-2026-10-07/


    #AINews #ArtificialIntelligence #Nvidia #ReflectionAI #OpenAI #AISafety #DataPrivacy #UKICO #Samsung #AIChips #OpenWeight #TechNews

    This episode includes AI-generated content.
    続きを読む 一部表示
    7 分