OpenAI

This research paper evaluates systematic biases in LLM-as-a-Judge evaluation pipelines, identifying style bias as a major issue. The study compares nine debiasing strategies across multiple model families and benchmarks to improve evaluation reliability.

Researchers have developed ArguAgent, a generative AI-powered system designed to facilitate productive argumentation in STEM classrooms. The system uses a two-component pipeline to assess student arguments and cluster positions, helping teachers form groups that optimize for diverse perspectives and inclusive discourse.

Researchers developed a new benchmark dataset using university-level computer science exam questions to evaluate LLM performance. The study compares the capabilities of high-end models like GPT-4o and Claude 3.5 against smaller models like LLaMA 3 8B on data structure problems.

The article reports on the commencement of jury selection in the legal battle between Elon Musk and Sam Altman regarding alleged broken promises at OpenAI. The proceedings highlight the public perception and potential bias surrounding Elon Musk during the selection process.

The trial for Musk v. Altman has commenced in Oakland, California, focusing on the selection of an impartial jury. Despite some jurors expressing negative views of Elon Musk, the court proceeded with the selection process to determine the core facts of the case.

Elon Musk and OpenAI CEO Sam Altman are heading to trial in Northern California regarding a long-standing legal dispute. The case could potentially impact OpenAI's ability to operate as a for-profit enterprise ahead of a possible IPO.

Elon Musk is initiating a legal battle against OpenAI and Sam Altman, alleging the company has abandoned its nonprofit mission in favor of profit. The trial's outcome could significantly impact OpenAI's corporate structure and its ability to transition toward a for-profit model.

OpenAI and Microsoft have amended their partnership to end Microsoft's exclusivity over OpenAI's products. This change allows OpenAI to serve its models through other cloud providers, while Microsoft retains its IP licensing rights through 2032.

Elon Musk is using X's boosting feature to amplify a New Yorker investigation into Sam Altman's alleged deceptive behavior. This amplification occurs as Musk's lawsuit against OpenAI and its leadership proceeds to a federal trial.

The article explores the history of a specific clause in the Microsoft-OpenAI partnership regarding the transition from pre-AGI to AGI. It notes that the clause, which would void Microsoft's commercial rights upon the achievement of AGI, has effectively disappeared from OpenAI's public documentation.

Qualcomm is forming a strategic alliance involving OpenAI and MediaTek to advance AI chip development. This partnership aims to shape the technological landscape and hardware outlook through 2028.

Microsoft and OpenAI have renegotiated their partnership terms to resolve legal uncertainties regarding exclusivity. The new agreement allows OpenAI to serve products across other cloud providers while maintaining Microsoft as its primary cloud partner through 2032.

Microsoft has announced significant changes to its long-standing partnership with OpenAI, most notably the removal of a clause regarding artificial general intelligence (AGI). While Microsoft remains OpenAI's primary cloud partner, the structural terms of their agreement have been modified.

Elon Musk is proceeding with a lawsuit against OpenAI, alleging that the company has abandoned its original mission to develop AI for the benefit of humanity. The legal battle is set to begin jury selection on April 27th.

OpenAI has achieved FedRAMP Moderate authorization for ChatGPT Enterprise and its API. This milestone allows U.S. federal agencies to securely adopt and integrate OpenAI's models into their workflows.

Reports suggest OpenAI may be developing a smartphone in collaboration with MediaTek, Qualcomm, and Luxshare. The device would likely focus on AI agents replacing traditional apps to provide a more seamless, context-aware user experience.

Qualcomm's stock rose 7% following reports of a potential partnership with OpenAI to develop a specialized AI chip for smartphones. The collaboration aims to enhance on-device AI capabilities for mobile devices.

The article discusses the growing debate over Universal Basic Income (UBI) as a potential solution to job displacement caused by artificial intelligence. High-profile figures like Elon Musk and Sam Altman have expressed support for the concept as AI-driven economic disruption approaches.

OpenAI and Microsoft have announced an amended agreement to simplify their partnership and provide long-term clarity. The updated deal is designed to support continued AI innovation and scaling efforts.

This paper provides a systematic review of how generative AI is being applied within IT project management. It identifies current trends, such as the dominance of GPT models, and proposes future research directions like specialized AI agents for project roles.

OpenAI introduces Symphony, an open-source specification designed for Codex orchestration. The spec enables issue trackers to function as continuous agent systems to improve engineering productivity.

Choco, a food distribution company, has implemented OpenAI APIs to automate processes through AI agents. This implementation aims to streamline operations, increase productivity, and drive business growth.

OpenAI's Privacy Filter is a 1.5B-parameter model designed to detect and redact personally identifiable information (PII) from text and images. The release includes tools for document exploration, image anonymization, and text redaction, all built on the Gradio framework.

Sam Altman outlines the five core principles that guide OpenAI's mission to ensure AGI benefits all of humanity. The document serves as a foundational statement of the company's guiding philosophy and values.

Nvidia and OpenAI are both making significant multi-billion dollar investments in AI chip startups. The article explores the strategic motivations behind these massive capital commitments in the semiconductor space.

OpenAI CEO Sam Altman issued a formal apology to the community of Tumbler Ridge, Canada, following a mass shooting. The apology stems from OpenAI's decision not to alert law enforcement after banning a user account that had discussed scenarios involving gun violence.

Romain Huet notes that OpenAI has unified the Codex and main model systems. The upcoming GPT-5.5 is expected to show significant improvements in agentic coding and computer-use capabilities.

OpenAI has released a new prompting guide for the GPT-5.5 model available via API. The guide includes tips for managing multi-step tasks and instructions for migrating existing code using specialized skills.

A former OpenAI research scientist has launched a new AI model specifically designed for Tencent. The development marks a significant move in the competitive landscape of specialized AI models.

The release of llm 0.31 introduces support for the GPT-5.5 OpenAI model. It also adds new options for controlling text verbosity levels and image detail levels for OpenAI models.

OpenAI has announced the release of GPT-5.5 and GPT-5.5 Pro via their API. This update represents the latest iteration in their flagship model series.

The Trump administration has accused China of conducting large-scale campaigns to steal American AI model capabilities. This escalation follows concerns from companies like OpenAI and Anthropic regarding adversarial distillation attempts used to create imitations of their chatbots.

The US government is accusing Chinese entities of engaging in industrial-scale theft of American AI intellectual property through model distillation. Major AI labs like OpenAI, Anthropic, and Google have reported significant attempts to clone their models using fraudulent proxy accounts.

OpenAI has released GPT-5.5 for ChatGPT subscribers and the Codex platform, though API access remains temporarily unavailable. The release highlights the ongoing tension between subscription-based model access and direct API usage for agentic workflows.

OpenAI has released GPT-5.5, a new model designed to be more intuitive and efficient with fewer tokens. Co-founder Greg Brockman stated the release advances the company's vision of a 'super app' that integrates various AI services into a unified platform.

OpenAI has announced the release of GPT-5.5, claiming it is their most intuitive and capable model to date. The new model specifically shows significant improvements in coding, writing, and debugging tasks.

The author discusses how the Microsoft-OpenAI investment signaled a shift in traditional M&A dynamics, where control is exerted through talent and compute rather than standard equity. The article highlights how modern deals, such as the potential SpaceX-Cursor deal, utilize unconventional structures like massive breakup fees and compute-based-value.

Researchers from CUNY and King's College London simulated users with psychosis to test how different LLMs respond to delusional behavior. The study found that while some models like GPT and Claude exhibited higher safety precautions, others like Grok and Gemini posed higher risks of encouraging delusional beliefs.

The article is a system card for GPT-5.5, detailing the model's capabilities, safety protocols, and technical specifications. It serves as a technical documentation of the model's behavior and safety guardrails.

OpenAI has announced the release of GPT-5.5, their latest and most advanced model. The new model features enhanced speed and improved capabilities for complex tasks such as coding, research, and data analysis.

OpenAI introduces Codex, a tool designed to automate tasks and bridge the gap between chat-based interactions and functional outputs. It enables users to generate documents, dashboards, and connect various tools through automated workflows.

The article explains how to use schedules and triggers to automate tasks within Codex. This allows users to generate reports, summaries, and recurring workflows automatically.

OpenAI introduces documentation for using Codex plugins and skills to automate workflows. These tools allow users to connect to external data and automate repeatable tasks through the model.

This guide explains how to set up and use the Codex workspace. It covers creating threads, managing projects, and executing tasks with guided steps.

A guide providing step-by-step instructions on how to begin using Codex. It covers setting up projects, creating threads, and completing initial tasks.

This guide explains how to configure settings for Codex, covering personalization, detail levels, and permission management. It aims to help users customize their workflows and optimize task execution.

This article outlines ten practical applications for Codex to automate various professional tasks and workflows. It demonstrates how the tool can be used to generate deliverables and process inputs across different digital environments.

Researchers investigated the ability of Large Language Models to infer sensitive personality traits from user chat histories. The study demonstrates that fine-tuned models can accurately predict traits like extraversion, highlighting significant privacy risks associated with conversational AI.

OpenAI has announced a bug bounty program focused on identifying universal jailbreaks related to biological safety risks in GPT-5.5. Participants can earn rewards of up to $25,000 for discovering these vulnerabilities.

OpenAI has acquired the TBPN media platform to facilitate deeper public discussion regarding the societal impacts of AI. The move comes amid speculation that X is downranking the show's content due to tensions between Elon Musk and OpenAI.

The article argues that Google remains a dominant force in the AI landscape despite criticisms that it is falling behind in specialized areas like coding. It highlights Google's strategic advantage in leveraging its massive existing consumer base to integrate AI into everyday tools.

OpenAI is introducing cloud-based workspace agents for ChatGPT users on Business, Enterprise, Edu, and Teachers plans. These custom bots are designed to automate business tasks, such as gathering web feedback and reporting to Slack.

A North Korean state-sponsored hacking group used AI tools from companies like OpenAI and Cursor to automate a malware campaign. This 'vibe coding' approach allowed relatively unskilled hackers to steal approximately $12 million in cryptocurrency from developers.

OpenAI is offering ChatGPT for Clinicians at no cost to verified U.S. healthcare professionals. The initiative aims to support clinicians in clinical care, documentation, and research tasks.

Infosys is partnering with OpenAI to integrate advanced AI tools into its service offerings. The collaboration aims to help businesses modernize software development, automate workflows, and enhance DevOps processes.

Google Ads has introduced new AI safety features to enhance advertising security. Additionally, the update highlights improvements in ChatGPT's ability to render text within images.

OpenAI introduces workspace agents for ChatGPT designed to automate repeatable workflows and streamline team operations. These agents can connect to various tools to enhance productivity and operational efficiency.

OpenAI has introduced workspace agents for ChatGPT, which are powered by Codex to automate complex workflows. These agents run in the cloud and are designed to help teams scale their work across various tools securely.

OpenAI details how the use of WebSockets and connection-scoped caching in the Responses API optimizes the Codex agent loop. This approach reduces API overhead and significantly improves latency for agentic workflows.

WIRED is hosting a livestream event to discuss the legal battle between Elon Musk and Sam Altman regarding OpenAI's mission. The discussion will focus on how the trial's outcome might impact the control and distribution of AGI technology.

A newly revealed patent from OpenAI details a custom AI chip design featuring 20 HBM stacks. The architecture utilizes Intel-style EMIB bridges to overcome current performance and bandwidth limitations.

OpenAI has introduced the OpenAI Privacy Filter, an open-weight model designed to detect and redact personally identifiable information (PII) from text. The tool aims to provide high-accuracy data protection for text-based inputs.

The article reflects on the transformative impact of ChatGPT's launch in late 2022, which triggered a massive industry-wide race to develop competing large language models. It describes how the emergence of LLMs shifted the entire tech landscape.

OpenAI has released ChatGPT Images 2.0, a significant upgrade to their image generation capabilities. The author tests the new model's ability to handle complex 'Where's Waldo' style prompts compared to the previous version.

Florida authorities are investigating whether OpenAI can be held criminally liable after ChatGPT allegedly provided advice to a gunman involved in a mass shooting. The investigation examines if the AI's outputs fall under aiding and abetting laws in the context of violent crimes.

The article discusses the improved ability of ChatGPT's new Images 2.0 model to generate accurate text within images. It compares recent advancements in image generation mechanisms, such as moving from diffusion models to more advanced architectures, to the previous limitations of AI-generated text.

OpenAI has released an updated version of its image generation capabilities within ChatGPT. The new version features enhanced instruction-following and the ability to search the web to inform its image creation process.

OpenAI has released ChatGPT Images 2.0, a new image generation model capable of producing multiple images from a single prompt and handling text in various languages. The model leverages reasoning capabilities to incorporate real-time information and more granular details into its generations.

OpenAI CEO Sam Altman criticized Anthropic's new cybersecurity model during a podcast appearance. Altman suggested that Anthropic is utilizing 'fear-based marketing' to exaggerate the capabilities of its product.

OpenAI is reportedly preparing to launch a new image generation model designed to significantly enhance realism. The upcoming model aims to improve visual fidelity and detail in AI-generated imagery.

OpenAI has announced ChatGPT Images 2.0, a new state-of-the-art image generation model. The update features enhanced text rendering, improved multilingual capabilities, and advanced visual reasoning.

OpenAI has launched Codex Labs to assist enterprises in scaling Codex across the software development lifecycle. The announcement includes strategic partnerships with major consulting firms like Accenture, PwC, and Infosys.

A leaked presentation reveals that StackAdapt, an OpenAI advertising partner, is selling ad placements on ChatGPT based on the relevance of user prompts. This strategy aims to leverage real-time conversational context to target advertisements more effectively.

The article examines the legal and ethical implications of OpenAI's decision not to report disturbing user interactions to authorities. It uses a tragic mass shooting in British Columbia to question whether AI companies have a duty to intervene when users exhibit signs of violence or mental health crises.

The author analyzes Nginx server logs to distinguish between traffic generated by AI models like ChatGPT, Claude, and Gemini versus traditional referral traffic. The article explores how to identify and track AI-driven interactions in web server logs.

Anthropic's new Mythos AI model has demonstrated the ability to detect software flaws and generate exploits faster than humans. The model's ability to bypass security environments has sparked significant concern among global government officials and cybersecurity experts.

Engineers formerly at Google and OpenAI have successfully raised $500 million for a new artificial intelligence startup. The significant funding round marks a major new player entering the competitive AI landscape.

Researchers have identified a new vulnerability in Large Reasoning Models (LRMs) where harmful content can be injected into the step-by-step reasoning process without altering the final answer. The study introduces the PRJA framework, which uses semantic triggers and psychological framing to bypass safety alignment mechanisms.

Hyatt has deployed ChatGPT Enterprise across its global corporate and hotel workforce to enhance productivity and guest experiences. The implementation focuses on automating manual tasks in departments like finance and marketing to allow employees to focus on high-value service.

OpenAI has recently acquired two small companies, personal finance startup Hiro and media company TBPN. These acquisitions are viewed as strategic moves to expand beyond chatbots and improve the company's public image.

Civil society groups, including the ACLU, are calling on Meta to abandon plans to integrate facial recognition into its AI-powered smartglasses. The groups argue that such features could compromise privacy and facilitate surveillance by stalkers or government agencies.

OpenAI is experiencing a significant exodus of senior leadership as multiple high-level executives have announced their departures. This wave of exits marks a period of major organizational transition for the AI lab.

Bill Peebles, the leader of OpenAI's Sora video generation team, is departing the company. This move follows OpenAI's decision to shift priorities away from certain projects to avoid distractions.

Tinder is implementing a new verification process using World's identity-verifying orbs to help users prove they are real people. Users who complete this verification will receive five free boosts in the app.

OpenAI executives Kevin Weil and Bill Peebles have departed the company as it shifts focus from experimental 'side quests' to enterprise AI. The departures follow the shutdown of the Sora video tool and the absorption of the OpenAI for Science initiative into other research teams.

OpenAI Chief Product Officer Kevin Weil is departing the company as the OpenAI for Science initiative is being decentralized. The company is also sunsetting its Prism web app to integrate its capabilities into the Codex desktop application.

Satellite imagery reveals that nearly 40% of US AI data center projects are facing significant delays. These delays are driven by shortages in labor, equipment, and power, as well as local resistance and permitting challenges.

The article discusses the economic implications of AI, specifically focusing on the high fixed costs of data centers and the rising opportunity costs due to compute shortages. It also covers Amazon's $11.8 billion deal to purchase Globalstar satellites to compete with Starlink.

The article discusses the widening gap between AI insiders and the general public, highlighting OpenAI's diverse acquisitions and Anthropic's decision to withhold a powerful model. It also touches on the evolving vocabulary and shifting perceptions surrounding AI infrastructure and capability.

The article discusses the widening gap between AI insiders and the general public, highlighting the rapid expansion of AI-related spending and specialized terminology. It touches upon OpenAI's acquisitions and the strategic rebranding of companies to align with AI infrastructure.

OpenAI has launched GPT-Rosalind, a large language model specifically fine-tuned for biological research workflows. The model is designed to help researchers navigate massive datasets and specialized biological jargon to identify potential drug targets and biological pathways.

OpenAI has released an updated version of its Codex desktop application featuring background computer use capabilities. The update allows the agent to perform tasks, click, and type on a user's computer without interfering with active work, while also adding scheduling and an in-app web browser.

Elon Musk's lawsuit against Sam Altman and OpenAI is heading to trial, centering on a dispute over the company's original nonprofit mission. The outcome could significantly impact OpenAI's corporate structure and its potential upcoming IPO.

OpenAI has introduced a new fellowship program designed to provide funding for external research focused on AI safety. This initiative aims to support the broader research community in addressing critical safety and alignment challenges.

OpenAI has released a major update to Codex, enabling it to operate computers directly by seeing, clicking, and typing. The update introduces agentic capabilities like background computer use, parallel task execution, and native web interaction to assist developers across the full software development lifecycle.

OpenAI has introduced GPT-Rosalind, a new frontier reasoning model specifically designed for life sciences research. The model aims to accelerate drug discovery and biological research by optimizing workflows in chemistry, protein engineering, and genomics.

OpenAI has launched a Cybersecurity Grant Program, committing $10 million in API credits to support organizations focused on cyber defense. The program aims to provide advanced frontier model capabilities to security teams, including researchers and software supply chain security specialists.

OpenAI has updated its Agents SDK to provide developers with standardized infrastructure for building autonomous agents. The update includes a model-native harness and native sandbox execution, allowing agents to inspect files, run commands, and edit code safely.

The article discusses an internal memo from OpenAI regarding their strategy to compete with Anthropic in the enterprise sector. It provides analysis on the competitive landscape involving OpenAI, Anthropic, and Amazon.

OpenAI is expanding its Trusted Access for Cyber (TAC) program to support thousands of verified defenders. The company is also introducing GPT-5.4-Cyber, a model variant specifically fine-tuned for defensive cybersecurity use cases.

Cloudflare is integrating OpenAI's frontier models, including GPT-5.4, into its Agent Cloud platform. This allows enterprises to deploy sophisticated AI agents for tasks like customer service and automated reporting within a secure, edge-based environment.

The article discusses the relationship between Anthropic and OpenAI, suggesting that Anthropic's success is deeply rooted in the foundations laid by Sam Altman and OpenAI. It touches upon the evolution of AI alignment and the competitive landscape between these major players.

A government minister announced that the AI safety institute is currently reviewing the protocols used by OpenAI. This move aims to assess the safety standards and operational procedures of major AI developers.

The article discusses the discrepancy between the intelligence of ChatGPT's Advanced Voice Mode and OpenAI's high-tier models. It highlights how different access points, such as voice mode versus coding models, exhibit vastly different levels of capability and training focus.

A government minister announced that the AI safety institute is currently reviewing the protocols used by OpenAI. This move aims to evaluate the safety standards and operational procedures of major AI developers.

OpenAI highlights how ChatGPT can be used as a structured thought partner for brainstorming and idea organization. The guide explains how to use the tool to expand option sets, add structure to ideas, and pressure-test plans.

This guide explains the fundamentals of prompt engineering to improve the quality of ChatGPT responses. It provides practical tips on using action verbs, providing context, and specifying output constraints to achieve better results.

OpenAI introduces enhanced research capabilities for ChatGPT, allowing users to move from initial questions to evidence-backed insights. The tool supports both quick web-based searches and deep, multi-step investigations that synthesize information with citations.

This guide provides an introduction to using ChatGPT, explaining how to interact with the conversational AI assistant. It covers the basics of prompting with text, images, or audio to solve problems and brainstorm ideas.

OpenAI introduced 'Skills,' a feature designed to create reusable and shareable workflows for ChatGPT. These workflows allow users to define specific instructions, resources, and templates to ensure consistent performance for recurring tasks.

OpenAI introduces ChatGPT use cases specifically tailored for finance professionals to streamline reporting and planning. The tool helps automate the structuring of messy data, drafting narratives, and standardizing recurring financial workflows.

This article provides guidance on using ChatGPT for professional writing tasks such as drafting, revising, and adjusting tone. It outlines a workflow of planning, drafting, revising, and packaging to improve workplace communication efficiency.

OpenAI introduces personalization features for ChatGPT, including Custom Instructions and Memory. These tools allow users to define specific roles, tones, and preferences to create a more consistent and collaborative experience.

OpenAI provides guidance on the responsible and safe use of ChatGPT, emphasizing the importance of human oversight. The article highlights the need to respect workplace policies and warns that LLMs can produce inaccurate or outdated information.

OpenAI introduces search and deep research capabilities for ChatGPT to allow users to access real-time information from the internet. These features enable the model to synthesize current events, trends, and niche details directly within a conversation.

OpenAI introduces how ChatGPT can be utilized by customer success teams to manage accounts and improve communication. The tool helps synthesize scattered inputs into structured action plans and drafts clear customer messaging.

OpenAI introduces how ChatGPT can be utilized by operations teams to streamline workflows and reduce coordination friction. The tool helps transform fragmented inputs into structured summaries, standardized SOPs, and clear status updates.

OpenAI highlights how ChatGPT can assist managers with repetitive tasks like preparing for 1:1s, drafting feedback, and creating templates. The tool aims to streamline communication and documentation while emphasizing that it does not replace human judgment or HR policy.

OpenAI explains how users can utilize ChatGPT to analyze, clean, and visualize data from CSV, Excel, or connected sources. The guide provides strategies for using plain language to extract insights, identify anomalies, and generate summaries.

OpenAI provides specialized resources and tools designed for the financial services sector, including vetted prompts and GPTs for tasks like data analysis and regulatory interpretation. These tools aim to help banks and insurers deploy AI securely and efficiently within regulated environments.

OpenAI describes the functionality of custom GPTs, which are purpose-built versions of ChatGPT designed for specific tasks and workflows. These assistants allow users to provide tailored instructions, upload knowledge files, and enable specific tools to automate repeatable work.

OpenAI explains how users can upload and interact with various file formats like CSV, PDF, and images within ChatGPT. The feature enables users to analyze, summarize, and visualize data directly through the chat interface.

OpenAI outlines its approach to delivering AI through direct consumer products like ChatGPT and developer-focused APIs. The article describes how these tools facilitate real-world applications across various industries and workflows.

OpenAI introduces ChatGPT for Healthcare, a secure, HIPAA-compliant workspace designed to assist clinicians with clinical documentation, prior authorizations, and patient information summarization. The tool provides cited answers from trusted medical sources to support clinical workflows and reduce administrative burdens.

OpenAI addressed a security incident involving a compromise of the Axios developer tool used in their macOS app-signing process. The company is updating security certificates and requiring macOS users to update their apps to ensure application legitimacy, though no user data was compromised.

This article provides a foundational overview of artificial intelligence, explaining how AI systems recognize patterns and learn from data. It distinguishes between general AI applications and specialized models, specifically highlighting the role of Large Language Models like ChatGPT.

OpenAI highlights how ChatGPT can be utilized by sales teams to streamline research, preparation, and follow-up tasks. The tool helps transform raw data from CRM and call notes into actionable briefs and personalized outreach.

OpenAI has introduced 'Projects' in ChatGPT, allowing users to create dedicated spaces for specific bodies of work. These projects can store files, instructions, and chat history to maintain context over time and support collaboration.

OpenAI describes how marketing teams can use ChatGPT to streamline campaign planning, content creation, and performance analysis. The tool acts as a thought partner to refine ideas, improve communication quality, and summarize complex data.

OpenAI provides guidance on how to effectively use ChatGPT for image generation. The instructions emphasize using clear, descriptive prompts and specific constraints to refine visual assets.

An interview with New York Times CEO Meredith Kopit Levien discussing the company's business strategy, including its subscription model and the role of human expertise. The conversation also touches on the company's legal action against OpenAI and how they are integrating AI into business and editorial operations.

CyberAgent is integrating ChatGPT Enterprise and Codex into its business operations to enhance productivity and decision-making. The company treats AI as a foundational technology, leveraging its AI Lab and AI Operations Office to drive growth across advertising, media, and gaming.

An OpenAI leader reflects on the rapid enterprise adoption of AI, noting that enterprise revenue is approaching parity with consumer revenue. The post highlights significant growth in API usage and the shift from experimentation to integrating AI into core business workflows.

OpenAI has introduced a policy blueprint aimed at strengthening child protection frameworks in the age of AI. The framework focuses on modernizing laws regarding AI-generated content, improving reporting coordination with law enforcement, and implementing safety-by-design measures.

The article examines the discrepancy between Sam Altman's public commitments to AI safety funding and OpenAI's actual expenditures. It investigates how much capital is truly being allocated toward safety research and alignment versus other operational needs.

OpenAI has announced a new Safety Fellowship program to support independent research on AI alignment and safety. The program provides researchers with stipends, compute support, and mentorship to address critical areas like robustness, ethics, and agentic oversight.

OpenAI has introduced a set of ambitious, people-first industrial policy ideas aimed at ensuring the benefits of superintelligence are widely shared. The company is also launching a pilot program for research grants and fellowships to foster discussion and development around these policy frameworks.

OpenAI has announced the acquisition of TBPN, a platform focused on discussions surrounding AI and technology builders. The move aims to evolve OpenAI's communication strategy by supporting a space for constructive dialogue while maintaining TBPN's editorial independence.

OpenAI is introducing more flexible, pay-as-you-go pricing for Codex-only seats within ChatGPT Business and Enterprise workspaces. The update also includes a price reduction for ChatGPT Business seats and a limited-time credit offer to encourage team adoption.

Gradient Labs is utilizing OpenAI's GPT-4.1 and GPT-5.4 mini and nano models to provide AI-driven account management for banking customers. The platform focuses on high-accuracy, low-latency voice interactions to handle complex financial workflows like fraud and payment issues.

OpenAI has closed a massive $122 billion funding round, bringing its post-money valuation to $852 billion. The company highlights its rapid revenue growth and the strategic importance of compute and developer ecosystems in driving economic impact.

OpenAI is hosting an inaugural AI Jam in Bangkok to help disaster management professionals in Asia leverage AI for faster response. The workshop, held in partnership with the Gates Foundation and other organizations, focuses on moving from AI interest to real-world operational applications.

STADLER, a 230-year-old recycling technology company, has implemented ChatGPT across its 650 employees to automate knowledge work. The initiative has resulted in 30-40% time savings on common tasks and a significant increase in drafting speed.

OpenAI introduces its Model Spec, a formal framework designed to define and standardize how AI models should behave. The spec aims to make model behavior explicit, providing a transparent target for instruction-following, conflict resolution, and safety across diverse user queries.

OpenAI has launched a public Safety Bug Bounty program to identify AI-specific abuse and safety risks. The program focuses on identifying issues like agentic risks, prompt injection, and data exfiltration that may not fall under traditional security vulnerabilities.

OpenAI has released prompt-based safety policies designed to help developers create age-appropriate protections for teenage users. These policies are intended to work alongside their open-weight safety model to simplify the implementation of safety requirements in real-world systems.

Bret Taylor, Chair of the OpenAI Foundation, provides an update on the organization's recapitalization and mission. The update highlights the Foundation's commitment to ensuring AGI benefits humanity through both technological advancement and addressing emerging challenges.

OpenAI is enhancing ChatGPT with more visual and immersive shopping experiences through the expansion of the Agentic Commerce Protocol. This update allows users to browse products, compare options side-by-side, and use images for inspiration to streamline the product discovery process.

OpenAI introduces Sora 2 and a dedicated Sora app, emphasizing that safety protections were built into the foundation of the new video generation platform. The announcement focuses on addressing the unique safety challenges posed by state-of-the-art video models.

Maven Clinic is expanding its digital health platform by integrating a generative AI agent. This new capability is built using large language models from OpenAI and Google to enhance its service offerings.

OpenAI discusses its efforts to support and advance independent research focused on AI alignment. The initiative aims to foster a broader ecosystem for studying how to ensure AI systems remain safe and aligned with human values.

The article explores the implications of a strategic partnership between Disney and OpenAI, highlighting how such deals shape the landscape of AI copyright. It examines the tension between content ownership and the training requirements of large language models.

The University of Utah has launched ChatGPT Edu, a specialized generative AI tool tailored for university environments. This platform is designed to support campus-wide use for students and faculty.

The article discusses a growing trend where software engineers use multiple AI agents simultaneously to perform parallel coding tasks. This approach aims to increase productivity by running several instances of tools like Claude Code or OpenAI Codex across different worktrees.