OpenAI has banned several ChatGPT accounts with suspected ties to Chinese government entities. These accounts were flagged for requesting proposals on monitoring social media conversations. OpenAI's public threat report indicates these requests violated their national security policy. The company also banned Chinese-language accounts involved in phishing and malware campaigns. These accounts sought assistance from ChatGPT in automating such activities. Furthermore, OpenAI identified and banned accounts linked to suspected Russian-speaking criminal groups. These groups used the chatbot to aid in developing malware. This action highlights concerns about the misuse of generative AI. The competition between the U.S. and China in AI development is intensifying. OpenAI's report underscores the safety challenges posed by these advanced technologies.

"Please, just stop sending me AI videos of Dad," the late Robin Williams' daughter Zelda wrote on Instagram.

OpenAI CEO Sam Altman revealed a vision at DevDay to transform the digital world beyond simple chatbots. The company is evolving into a comprehensive computing platform, moving from model provision to building an entire ecosystem. ChatGPT is being reinvented as a platform with an Apps SDK, allowing developers to build and distribute applications directly within the chat interface. These apps can offer interactive and personalized experiences, reaching OpenAI's vast user base. Beyond ChatGPT, OpenAI is empowering AI agents with Agent Kit, enabling autonomous AI workers to automate complex tasks and business processes. The company is also significantly enhancing its AI coding agent, Codex, powered by GPT-5, which can now autonomously write, review, and integrate code. A significant, unannounced collaboration with legendary designer Jony Ive was revealed, focusing on developing new AI-centric hardware. Ive believes current products are inadequate for advanced AI and aims to reset humanity's relationship with technology. A critical constraint for this ambitious future is the massive demand for computing power, driving OpenAI's substantial infrastructure investments. The company views AI as the future fundamental driver of economic growth. OpenAI is not just creating AI; it's constructing the future environment for AI, encompassing intelligent applications, agents, and hardware.

Google's DeepMind has released Gemini 2.5 Pro Computer Use, an AI model designed to act as a virtual agent on the web. This new model can navigate websites, fill forms, and perform actions on behalf of users, similar to offerings from OpenAI and Anthropic. Google CEO Sundar Pichai highlighted its importance in developing general-purpose AI agents. While not directly available to consumers, it's accessible through the Browserbase platform and the Gemini API for developers. The model builds on Gemini 2.5 Pro's capabilities, with a focus on interacting with user interfaces. It allows AI systems to operate visually and functionally, unlike API-dependent models. Early tests show success in navigating websites and completing tasks, though it lacks direct file system access of competitors. Google claims Gemini 2.5 Computer Use leads in interface control benchmarks and offers lower latency. The model operates in an interaction loop, analyzing screenshots and user prompts to recommend actions. Safety measures include per-step inspection and developer-defined instructions. It supports various UI actions like clicking and typing, with normalized screen coordinates. Pricing is similar to Gemini 2.5 Pro, but Computer Use is exclusively a paid tier offering. Data from paid tier usage does not improve Google products, unlike the free tier of Gemini 2.5 Pro.

Starting today, we’re bringing AI Mode in Google Search to more people around the world, launching in more than 35 new languages and over 40 new countries and territorie…

Marketers who ignore this truth are sheep walking into a slaughterhouse. Those who embrace it become wolves. Each ad, each offer, each suggestion is an invisible push toward a predetermined outcome.

AMD's unusual deal to finance OpenAI's chip purchases could grant the AI model maker up to $100 billion, analyst estimates.

Startups are increasingly leveraging agentic AI to automate workflows and solve complex business challenges. Google Cloud is launching a technical guide specifically for startups to navigate this new landscape of AI agents. AI agents combine AI models and tools to perform actions, offering proactive problem-solving capabilities. This guide provides a roadmap for building AI agents, addressing the associated technical complexities. Startups can choose to build, use, or integrate agents, utilizing Google Cloud's flexible ecosystem. The guide outlines four key steps for building an agent using the Agent Development Kit (ADK). These steps involve defining the agent's identity, providing instructions, equipping it with tools, and mastering its lifecycle through testing and deployment. The guide emphasizes the importance of AgentOps for ensuring quality and reliability. Wotter utilizes agentic methods and Gemini 2.5 Flash for employee engagement insights, while Harvey employs Gemini 2.5 Pro for complex legal document analysis. The guide, combined with the Google for Startups Cloud Program, aims to accelerate startups' AI adoption.

U.S. GDP growth in early 2025 was overwhelmingly fueled by investments in data centers and information processing technology. Without these tech-specific investments, GDP growth would have been minimal, only 0.1% annually. While accounting for a small portion of overall GDP, these technology investments were responsible for the vast majority of its growth. Major tech companies have invested billions in data center expansion, with AI data-center buildout alone contributing significantly to GDP, even surpassing consumer spending's contribution to growth.

The hot-handed bitcoin miner turned high-performance computing play could raise as much as $1 billion in the note sale.

The app, which lets you create mini web apps using text prompts, is now available in Canada, India, Japan, South Korea, Vietnam, Indonesia, Brazil, Singapore, Colombia, El Salvador, Costa Rica, Panamá, Honduras, Argentina, and Pakistan.

MrBeast, also known as Jimmy Donaldson, has expressed concerns about the future of the creator economy due to advancements in AI video technology. He worries that as AI-generated videos become indistinguishable from human-created content, it will significantly impact creators who rely on YouTube for their income. This concern is amplified by recent releases from companies like OpenAI with its Sora platform and Meta with its Vibes platform, both capable of generating realistic AI videos. Donaldson views these developments as potentially "scary times" for the millions of content creators on platforms like YouTube.

Otter is launching a suite of enterprise tools to help companies use its tech to create a central knowledge base.

Google AI Plus — our newest AI plan — expands to 36 more countries.

Today we are releasing the Gemini 2.5 Computer Use model via the API, which outperforms leading alternatives at browser and mobile tasks.

Conversational AI has long promised human-like assistants, but reliably completing tasks outside of chat remains a significant challenge, with even top models failing frequently on benchmarks. New York-based Augmented Intelligence (AUI) Inc. believes its new foundation model, Apollo-1, offers a solution through stateful neuro-symbolic reasoning. This hybrid architecture aims to guarantee consistent, policy-compliant outcomes, unlike purely generative LLMs that produce probable but not certain results. While LLMs excel at open-ended dialogue, Apollo-1 is designed for task-oriented conversations requiring certainty and predictable behavior. AUI's approach involves a closed reasoning loop that translates natural language into a symbolic state, maintaining and acting upon it to achieve deterministic task completion. Apollo-1 is intended as a domain-agnostic foundation model configurable via a "behavioral contract" System Prompt. Developed over eight years, the model separates procedural and descriptive knowledge from task-oriented conversations. Benchmarks show Apollo-1 significantly outperforms existing LLMs in task completion rates across various domains. AUI positions Apollo-1 as a complementary counterpart to LLMs, completing the spectrum of conversational AI. The model is currently in limited pilots with Fortune 500 companies and plans a general release in November 2025 with expanded capabilities and API access. Apollo-1's core promise is to make AI that businesses can trust to act reliably, potentially bridging the gap between talking and doing.

As AI models grow larger and cheaper to run, Earth’s data centers face limits in energy, cooling, and community tolerance. Space offers continuous solar power, zero local externalities, and radiation-based cooling—making it an ideal compute environment. Yet orbital infrastructure comes with its own challenges: launch logistics, maintenance, and governance. The likely path forward starts small—with AI models running alongside satellites—before scaling to full orbital training campuses. It’s a glimpse of a future where most AI computing happens beyond Earth’s atmosphere.

Google now celebrates five Nobel laureates, including three prizes in the past two years.

You can now access apps like Spotify, Canva, and Zillow inside ChatGPT. This could turn the chatbot into a powerful new marketing channel.

Learn how to effectively review AI-generated .NET code with practical strategies for maintaining quality, enforcing standards, and boosting team productivity through thoughtful code review practices.

Kids have always played with and talked to stuffed animals. But now their toys can talk back, thanks to a wave of companies that are fitting children’s playthings with chatbots and voice assistants.  It’s a trend that has particularly taken off in China: A recent report by the Shenzhen Toy Industry Association and JD.com predicts…

OpenAI has committed approximately $1 trillion in deals for computing power this year, including agreements with AMD, Nvidia, Oracle, and CoreWeave. These commitments far exceed OpenAI's current revenue, raising concerns about its ability to finance these substantial obligations. The deals secure over 20 gigawatts of computing capacity for the next decade, with estimates suggesting each gigawatt costs around $50 billion to deploy. These agreements bind major tech companies to OpenAI's future profitability and its capacity to meet these mounting financial requirements.

Anthropic's next big bet is India — one of its fastest-growing markets worldwide.

Anthropic's Claude large language model family will be incorporated into some of IBM's software development products.

Crunch Lab raises $5M from Galaxy Ventures to build decentralized AI network with 10,000+ engineers, delivering 17% accuracy gains for ADIA Lab.