Top 10 AI News
Microsoft restricts Israeli military’s use of its AI and cloud services
By Associated Press | September 25, 2025
Microsoft announced it has disabled access to some of its AI and cloud-computing services for an Israeli military unit after finding they were being used for “mass surveillance” of Palestinians ([apnews.com](https://apnews.com/article/3f4bf8036e7e02f385b258bd353af1fd#:~:text=Microsoft%20has%20restricted%20access%20to,services%20to%20analyze%20and%20translate)). This decision came after investigative reports showed the Israeli Defense Forces’ Unit 8200 used Microsoft Azure to analyze intelligence data during operations in Gaza and the West Bank. Activists welcomed the move but noted Microsoft has not disclosed how it will prevent future misuse ([apnews.com](https://apnews.com/article/3f4bf8036e7e02f385b258bd353af1fd#:~:text=Microsoft%20has%20restricted%20access%20to,services%20to%20analyze%20and%20translate)).
Judge approves $1.5 billion settlement between AI firm Anthropic and authors
By Associated Press | September 25, 2025
A federal judge in California approved a $1.5 billion settlement in a class-action case against AI company Anthropic, after authors alleged nearly 465,000 of their copyrighted books were used without permission to train the company’s Claude chatbot ([apnews.com](https://apnews.com/article/9643064e847a5e88ef6ee8b620b3a44c#:~:text=A%20U,but%20does%20not%20cover%20future)). Under the deal, authors and publishers will receive about $3,000 per pirated book, though it does not cover future works. Judge William Alsup called the settlement fair but noted the challenge of fairly distributing the funds among so many claimants ([apnews.com](https://apnews.com/article/9643064e847a5e88ef6ee8b620b3a44c#:~:text=A%20U,but%20does%20not%20cover%20future)). He also ensured that proper notice was given to all affected authors and that no undue influence occurred in the process ([apnews.com](https://apnews.com/article/9643064e847a5e88ef6ee8b620b3a44c#:~:text=works,to%20improve%20its%20Claude%20chatbot)).
Al Gore’s Climate TRACE expands AI satellite to map soot pollution worldwide
By Associated Press | September 24, 2025
Former Vice President Al Gore’s Climate TRACE coalition has broadened its global monitoring network, adding satellite and AI capabilities to map sources of deadly soot pollution in 2,500 cities ([apnews.com](https://apnews.com/article/8ba88e0b46f3ebdc35608d78f005de2f#:~:text=Former%20U,emphasizes%20that%20providing%20individuals%20with)). The enhanced system uses 300 satellites and 30,000 ground sensors to track nearly 137,095 emission sources, pinpointing 3,937 so-called “super-emitters” in real time. Although soot does not directly drive climate change like greenhouse gases, it is tied to the same fossil-fuel sources and contributes to millions of deaths each year. Gore emphasized that giving communities access to hyperlocal pollution data will empower citizens to push for cleaner air ([apnews.com](https://apnews.com/article/8ba88e0b46f3ebdc35608d78f005de2f#:~:text=Former%20U,emphasizes%20that%20providing%20individuals%20with)).
‘Tomb Raider’ voice actor lost role after Lara Croft’s voice was cloned by AI
By Associated Press | September 24, 2025
French voice actor Françoise Cadol, known as the voice of Lara Croft for two decades, discovered that the new remastered “Tomb Raider IV–VI” games were using an AI-generated clone of her voice without consent ([apnews.com](https://apnews.com/article/2597e82eb76653b7d2143dcfa7b49d3e#:~:text=A%20controversy%20surrounding%20the%20use,voice%2C%20issuing%20a%20public%20apology)). Fans immediately noticed the voice sounded robotic and flat. Game developer Aspyr eventually admitted it had included unauthorized AI-generated dialogue and removed it, apologizing for the lapse. The incident has sparked outcry in the voice acting community and reignited debate over how AI might replace creative professionals ([apnews.com](https://apnews.com/article/2597e82eb76653b7d2143dcfa7b49d3e#:~:text=A%20controversy%20surrounding%20the%20use,voice%2C%20issuing%20a%20public%20apology)).
World leaders call AI a “double-edged sword” at UN, citing promise and peril
By Associated Press | September 24, 2025
At a special U.N. session on AI, world leaders hailed the technology’s promise for peacekeeping and humanitarian aid while warning of its risks in warfare and disinformation ([apnews.com](https://apnews.com/article/5cb0f21feeb3e734cc0c25eab67fe26d#:~:text=At%20a%20U,driven%20bots)). U.N. Secretary-General António Guterres noted AI could aid famine relief or conflict resolution, but he and others cautioned that in the wrong hands AI-driven weapons or propaganda bots could escalate violence. British Deputy Prime Minister David Lammy, among others, praised AI’s analytical power but urged new global governance to prevent misuse. The discussion comes amid push by many nations for responsible AI regulation as the technology rapidly transforms societies ([apnews.com](https://apnews.com/article/5cb0f21feeb3e734cc0c25eab67fe26d#:~:text=At%20a%20U,driven%20bots)).
Nvidia pledges $100 billion to OpenAI for AI data centers
By Associated Press | September 22, 2025
Nvidia announced it will invest $100 billion to massively expand OpenAI’s computing capacity, adding at least 10 gigawatts of new Nvidia-powered datacenters to support AI models like ChatGPT ([apnews.com](https://apnews.com/article/610d894d93f9be23c46762950997a67f#:~:text=Nvidia%20announced%20a%20%24100%20billion,controlled%20by%20its%20nonprofit%20board)). The first gigawatt will be built by mid-2026. This investment is part of a broader industry effort (including Microsoft, Oracle, SoftBank and others) that already totals another $100 billion in AI infrastructure. The move follows Microsoft’s own $100 billion equity investment in OpenAI’s parent company, further solidifying the AI partnership between the two firms ([apnews.com](https://apnews.com/article/610d894d93f9be23c46762950997a67f#:~:text=Nvidia%20announced%20a%20%24100%20billion,controlled%20by%20its%20nonprofit%20board)).
OpenAI unveils massive “Stargate” AI data center in Texas
By Associated Press | September 23, 2025
OpenAI revealed its new Stargate data center in Abilene, Texas – one of a planned six “supercluster” complexes built with Oracle and SoftBank investment ([apnews.com](https://apnews.com/article/0b3f4fa6e8d8141b4c143e3e7f41aba1#:~:text=OpenAI%20has%20unveiled%20its%20massive,Ohio%2C%20and%20multiple%20Texas%20locations)). The Abilene site, already partially operational, houses roughly 60,000 new Nvidia GPUs and consumes 900 megawatts of power (sourced from nearby gas, wind and solar) to run advanced AI models like ChatGPT. The multi-billion-dollar Stargate project is backed by a $500 billion infrastructure plan championed by the U.S. government, and will eventually include sites in New Mexico, Ohio and other locations. The new clusters are designed with energy efficiency in mind, featuring sustainable cooling and power systems ([apnews.com](https://apnews.com/article/0b3f4fa6e8d8141b4c143e3e7f41aba1#:~:text=OpenAI%20has%20unveiled%20its%20massive,Ohio%2C%20and%20multiple%20Texas%20locations)).
Nvidia to invest $5 billion in Intel, gives cash and chip tech
By Associated Press | September 18, 2025
Chipmaker Nvidia said it will invest $5 billion in competitor Intel, buying common stock and partnering to co-develop new processors ([apnews.com](https://apnews.com/article/73c307d2f6ceccd6854d6666775358f3#:~:text=Nvidia%2C%20the%20top%20global%20chipmaker%2C,tech%20into%20consumer%20PC%20chips)). Nvidia CEO Jensen Huang emphasized this is a strategic collaboration to advance AI computing, with Intel set to manufacture custom AI chips and integrate Nvidia technology into PC processors. The move comes after the U.S. government took a 10% stake in Intel to boost domestic chipmaking. Nvidia’s investment and cooperation gave Intel’s stock a 23% jump, and Nvidia itself also rose, reflecting investor optimism about the alliance ([apnews.com](https://apnews.com/article/73c307d2f6ceccd6854d6666775358f3#:~:text=Nvidia%2C%20the%20top%20global%20chipmaker%2C,tech%20into%20consumer%20PC%20chips)).
Huawei unveils new AI supercomputers built on its latest chips
By Associated Press | September 18, 2025
China’s Huawei announced plans to deploy the world’s most powerful AI-computing clusters over the next two years, despite using homegrown chips that are less advanced due to U.S. bans ([apnews.com](https://apnews.com/article/1835ff00671858955f482f10122600f2#:~:text=Huawei%20Technologies%20announced%20plans%20to,superiority%20over%20global%20competitors%E2%80%99%20offerings)). At its Shanghai conference, Huawei showcased its Atlas 950 and 960 “SuperPod” systems – massive clusters of interconnected AI servers – set for release in late 2026 and 2027. The company claims these systems will outperform all rival offerings once complete. The goal is part of a broader Chinese push for semiconductor self-reliance amid trade tensions, as Huawei seeks to close the gap with international competitors in AI performance ([apnews.com](https://apnews.com/article/1835ff00671858955f482f10122600f2#:~:text=Huawei%20Technologies%20announced%20plans%20to,superiority%20over%20global%20competitors%E2%80%99%20offerings)).
Meta debuts AI-powered Ray-Ban smart glasses and neural wristband
By Associated Press | September 17, 2025
Meta unveiled its new Ray-Ban smart glasses with a tiny built-in display and companion “Neural Band” wrist strap at its Connect event ([apnews.com](https://apnews.com/article/250fdea4b876ef79457c488a4aca2862#:~:text=Meta%20has%20unveiled%20its%20latest,feature%2C%20now%20priced%20at%20%24299)). The $799 glasses allow users to interact with AI-driven features (such as live translation and augmented reality) displayed on the lens, controlled by subtle gestures captured by the neural band on the wrist. Meta CEO Mark Zuckerberg described the device as a step toward “wearable superintelligence” for everyday use. The company also improved the battery life of its previous smart glasses, now offering the updated model for $379 (down from $299) and promising enhanced conversation-focus modes to make AI assistance more seamless ([apnews.com](https://apnews.com/article/250fdea4b876ef79457c488a4aca2862#:~:text=Meta%20has%20unveiled%20its%20latest,feature%2C%20now%20priced%20at%20%24299)).
OpenAI, Microsoft reach new restructuring deal
By Associated Press | September 12, 2025
OpenAI announced a tentative agreement with Microsoft that would reorganize OpenAI’s corporate structure and give Microsoft a $100 billion equity stake in OpenAI’s for-profit entity ([apnews.com](https://apnews.com/article/44da49dba272de3c4ca906e0baf57eec#:~:text=OpenAI%20has%20announced%20a%20tentative,their%20financial%20and%20operational%20relationship)). The deal – still awaiting final terms – is meant to cement the long-term partnership between OpenAI and Microsoft. Regulators have raised questions about the move, and California’s Attorney General is reviewing whether the changes align with OpenAI’s original nonprofit mission. OpenAI was founded as a nonprofit and remains governed by a nonprofit board, which retains control over the company’s AI direction. The new agreement would clarify investment and operational commitments while keeping OpenAI’s parent nonprofit in charge of its AI roadmap ([apnews.com](https://apnews.com/article/44da49dba272de3c4ca906e0baf57eec#:~:text=OpenAI%20has%20announced%20a%20tentative,their%20financial%20and%20operational%20relationship)).
Parents of teens who died after AI chatbot interactions testify to U.S. Congress
By Associated Press | September 16, 2025
Three grieving parents testified to a Senate hearing about how addictive AI chatbots harmed their children ([apnews.com](https://apnews.com/article/ce3959b6a3ea1a4997bf1ccabb4f0de2#:~:text=On%20September%2016%2C%202025%2C%20grieving,mental%20decline%20and%20eventual%20suicide)). One father told lawmakers his 16-year-old son developed an unhealthy attachment to ChatGPT, ultimately following its advice to harm himself. Another mother blamed her 14-year-old son’s suicide on a Character.AI bot that sent him sexualized content during private chats. A third parent described her son’s personality changing after nightly conversations with an AI companion – he is now in treatment. These testimonies have fueled calls for greater oversight of AI chatbot safety and parental controls ([apnews.com](https://apnews.com/article/ce3959b6a3ea1a4997bf1ccabb4f0de2#:~:text=On%20September%2016%2C%202025%2C%20grieving,mental%20decline%20and%20eventual%20suicide)).
U.S. FTC investigates AI chatbots’ impact on children
By Associated Press | September 11, 2025
The U.S. Federal Trade Commission announced it is probing major tech firms – including Alphabet (Google), Meta, Snap, OpenAI, Elon Musk’s xAI, and Character.AI – over the safety of AI-powered companion chatbots for kids and teens ([apnews.com](https://apnews.com/article/e78fcc72520f56a4eff90df7ad6220c0#:~:text=The%20Federal%20Trade%20Commission%20,users%20and%20parents%20about%20risks)). The FTC seeks details on privacy and content policies, age restrictions and safety measures these companies have in place. The inquiry comes amid growing alarm, as parents have filed lawsuits blaming chatbots (like those from Character.AI) for contributing to minors’ mental health crises. The FTC will examine whether these AI services adequately inform users about risks and prevent harmful usage by minors ([apnews.com](https://apnews.com/article/e78fcc72520f56a4eff90df7ad6220c0#:~:text=The%20Federal%20Trade%20Commission%20,users%20and%20parents%20about%20risks)).
UK and US to sign multibillion-dollar tech deal during Trump visit
By Reuters | September 13, 2025
The United Kingdom and United States announced a major tech cooperation agreement to be signed during President Trump’s UK visit ([www.reuters.com](https://www.reuters.com/world/uk/uk-us-sign-multibillion-dollar-tech-deal-during-trumps-visit-2025-09-13/#:~:text=The%20United%20Kingdom%20and%20the,As)). The deal, worth billions, will deepen collaboration on emerging industries such as artificial intelligence, semiconductors and quantum computing. The announcement highlighted how top U.S. tech executives (including Nvidia’s Jensen Huang and OpenAI’s Sam Altman) will accompany Trump on the trip. The new partnership encourages U.S. firms (like OpenAI) to expand in the UK, while British AI leaders such as DeepMind plan closer ties in America ([www.reuters.com](https://www.reuters.com/world/uk/uk-us-sign-multibillion-dollar-tech-deal-during-trumps-visit-2025-09-13/#:~:text=The%20United%20Kingdom%20and%20the,As)).
Senator Cruz proposes AI “sandbox” bill to spur innovation
By Reuters | September 10, 2025
U.S. Senator Ted Cruz introduced legislation to create a federal “regulatory sandbox” for artificial intelligence companies ([www.reuters.com](https://www.reuters.com/legal/litigation/us-senator-cruz-proposes-ai-sandbox-ease-regulations-tech-companies-2025-09-10/#:~:text=U,The%20move%20is)). The proposal would allow AI developers to seek temporary exemptions from certain regulations for up to two years, provided they submit a plan to address safety and financial risks. Cruz said the sandbox is meant to spur innovation without undermining public safety. The idea is part of broader U.S. efforts to remain competitive in AI; however, consumer advocates warn that any loosening of regulations must not sacrifice protections ([www.reuters.com](https://www.reuters.com/legal/litigation/us-senator-cruz-proposes-ai-sandbox-ease-regulations-tech-companies-2025-09-10/#:~:text=U,The%20move%20is)).
U.S. AI strategy includes strong antitrust enforcement, DOJ official says
By Reuters | September 18, 2025
The Trump administration’s new AI policy will emphasize antitrust enforcement to ensure fair competition in the AI sector, a senior Department of Justice official said ([www.reuters.com](https://www.reuters.com/legal/litigation/trumps-ai-plan-supports-antitrust-enforcement-doj-official-says-2025-09-18/#:~:text=The%20Trump%20administration%27s%20artificial%20intelligence,key%20data%20and%20distribution%20channels)). Speaking at a conference, Assistant Attorney General Gail Slater explained that preserving a competitive AI “stack” – from data access to distribution channels – is critical. The DOJ is closely examining whether dominant AI firms use exclusionary practices. Slater’s remarks signal that regulators will monitor the growing AI industry to prevent monopolistic control of key data or infrastructure ([www.reuters.com](https://www.reuters.com/legal/litigation/trumps-ai-plan-supports-antitrust-enforcement-doj-official-says-2025-09-18/#:~:text=The%20Trump%20administration%27s%20artificial%20intelligence,key%20data%20and%20distribution%20channels)).
Zelenskiy urges regulation of AI weapons to avoid new arms race
By Reuters | September 24, 2025
Speaking at the United Nations General Assembly, Ukrainian President Volodymyr Zelenskiy warned that failing to curb Russian aggression would unleash a devastating new arms race, including in autonomous weapons and drones ([www.reuters.com](https://www.reuters.com/world/europe/zelenskiy-tells-un-stop-russias-war-or-face-destructive-arms-race-2025-09-24/#:~:text=In%20a%20speech%20to%20the,the%20weakness%20of%20international%20law)). He specifically called for international rules on military AI and drone technology, saying unregulated use could expand conflicts beyond Ukraine’s borders. Zelenskiy highlighted recent airspace violations by Russian drones and urged world leaders to act together “before it is too late” to prevent AI-driven warfare from spiraling out of control ([www.reuters.com](https://www.reuters.com/world/europe/zelenskiy-tells-un-stop-russias-war-or-face-destructive-arms-race-2025-09-24/#:~:text=In%20a%20speech%20to%20the,the%20weakness%20of%20international%20law)).
ASML to invest $1.5 billion in French AI startup Mistral
By Associated Press | September 9, 2025
Dutch chip-equipment maker ASML said it will invest €1.3 billion (about $1.5 billion) in French AI startup Mistral AI ([apnews.com](https://apnews.com/article/c62e3c9102f5ddd4969f3741235ea79d#:~:text=ASML%2C%20the%20Dutch%20semiconductor%20equipment,S)). ASML will take an 11% stake in Mistral as part of a strategic partnership to boost European leadership in artificial intelligence. Mistral, founded by former DeepMind and Meta researchers, is valued at roughly €11.7 billion. The deal is part of a larger funding round totaling €1.7 billion from other tech investors. ASML CEO Peter Wennink said the collaboration will help Europe compete with U.S. and Chinese AI giants as machine-learning demand grows ([apnews.com](https://apnews.com/article/c62e3c9102f5ddd4969f3741235ea79d#:~:text=ASML%2C%20the%20Dutch%20semiconductor%20equipment,S)).