OpenAI's Codex app version 26.527 brings two significant Windows upgrades. Computer Use now works on Windows, enabling Codex to operate desktop applications by seeing, clicking, and typing directly in the foreground. Remote control support has also been extended to Windows devices, allowing users to access and control their Windows machines remotely through the ChatGPT mobile app on either iOS or Android. Together, these additions make Codex a more capable tool for Windows-based automation and remote workflows.

Perplexity has expanded its Computer AI agent into Microsoft Word, Excel, PowerPoint, and Outlook via native add-ins available on Microsoft Marketplace. The integration removes the copy-paste workflow that nearly 40% of Computer users already relied on when generating Microsoft-format outputs. Users can draft documents, build financial models, design presentations, and compose emails from a side panel, with data pulled from the web, SharePoint, FactSet, and Snowflake. Every figure is cited. The feature is available to Pro, Max, and Enterprise subscribers.

Anthropic has raised $65 billion in a Series H funding round led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, valuing the company at $965 billion post-money. The round includes $15 billion in previously committed hyperscaler investments, including $5 billion from Amazon. Anthropic's run-rate revenue crossed $47 billion earlier this month. Funds will advance safety and interpretability research, expand compute, and scale products. Strategic infrastructure partners Micron, Samsung, and SK hynix also joined the round.

Microsoft has redesigned the Microsoft 365 Copilot app and its in-app experience across Word, Excel, PowerPoint, and Outlook. The updated app loads more than 50% faster and improves complex chat response times by 10%. A new intelligence layer called Work IQ draws on emails, files, chats, and meetings to adapt responses. Since rolling out the redesign, Copilot usage rose 27% in Word, 33% in Excel, 43% in PowerPoint, and 30% in Outlook.

OpenAI has launched Rosalind Biodefense, a program giving trusted developers sponsored access to GPT-Rosalind, its frontier reasoning model for life sciences, to build biodefense and pandemic preparedness tools. OpenAI is also expanding GPT-Rosalind access to select U.S. government and allied partners for public health missions. Early partners include Lawrence Livermore National Laboratory, Johns Hopkins Applied Physics Laboratory, the Coalition for Epidemic Preparedness Innovations supporting its 100 Days Mission, and Fourth Eon Biosecurity, which focuses on DNA synthesis screening.

OpenAI has published a playbook for conducting trustworthy third-party evaluations of frontier AI models, emphasizing that modern models require more than simple prompt-response testing. The guide highlights the critical role of the "harness" — the surrounding setup including tools, scaffolding, and context — which can significantly alter measured performance. OpenAI recommends evaluators specify claims being tested, document harness choices, and account for issues like reward hacking, sandbagging, and contamination. For example, GPT-5.5 cyber range scores improved materially when harness compaction was enabled.

Anthropic has launched dynamic workflows in Claude Code, now in research preview, enabling Claude to write orchestration scripts that spin up tens to hundreds of parallel subagents in a single session. Designed for large-scale engineering tasks like codebase-wide bug hunts, framework migrations, and security audits, the feature is available via the CLI, Desktop, VS Code extension, and API, including Amazon Bedrock, Vertex AI, and Microsoft Foundry, for Max, Team, and Enterprise plans. Developer Jarred Sumner used it to port Bun from Zig to Rust—roughly 750,000 lines—in eleven days.

ElevenLabs has launched Dubbing v2, a new dubbing model designed to preserve the emotion and performance of original content across multiple languages. The company describes it as a revolutionary advancement, marking the first time dubbed audio can carry over the emotional nuance of a source performance into every target language. This release positions ElevenLabs as a stronger competitor in AI-powered localization and content translation for global media distribution.

Amazon MGM Studios has launched the GenAI Creators' Fund, partnering with Amazon Web Services to build what they call the industry's first purpose-built AI production platform for visual storytelling, dubbed Project Nara. The fund has greenlit three animated Prime Video series: Punky Duck from Book of Life director Jorge R. Gutierrez, Love, Diana Music Hunters from former Nickelodeon president Albie Hecht, and Cupcake & Friends from BuzzFeed Studios. Executives emphasize the initiative empowers human creators rather than replacing them.

Microsoft is expanding protections against non-consensual intimate imagery (NCII) across its services, including Teams Free, OneDrive, and Xbox. The company has strengthened its reporting process with a clearer form covering both real and AI-generated images, and expanded use of StopNCII.org hashes—piloted on Bing since September 2024—to detect and remove known NCII. Microsoft also partnered with UK NGO Childnet to release educational materials on AI misuse for teens in the UK, Singapore, South Korea, and Japan.

OpenAI has outlined its 2026 election safeguards, building on 2024 efforts. ChatGPT will provide live vote counts from the Associated Press during US and Brazil elections this fall, and will partner with Democracy Works to display voting and registration information. OpenAI is offering its Codex Security and Trusted Access for Cyber programs to registered voting system manufacturers, and has partnered with Google's SynthID to watermark AI-generated images. A public verification tool at openai.com/verify lets users check if images were AI-generated.

YouTube is making AI content labels more prominent and adding automatic detection starting May 2026. For long-form videos, the label moves below the video player above the description; for Shorts, it appears as an overlay. If creators don't disclose AI use but YouTube's systems detect significant photorealistic AI content, a label will be applied automatically. Creators can dispute incorrect labels in YouTube Studio, though labels are permanent for content made with YouTube's own tools like Veo or Dream Screen.

Erin Brockovich, the environmental activist portrayed by Julia Roberts, has launched the Brockovich AI Data Center Reporting website, a crowdsourced map tracking operational and proposed AI data centers across the US. The site has received 2,716 reports, with the most coming from Texas, including 297 from Sulfur Springs, where MSB Global plans a gigawatt facility across 1,600 acres. Community concerns center on water usage, electricity costs, and health impacts from air and noise pollution.

The New York Times Tech Guild, representing around 700 tech employees, has filed grievances and unfair labor practice charges against the Times, alleging management violated their collective bargaining agreement by using two AI tools to monitor worker performance. The tools, DX, an engineering productivity tracker, and Glean, an internal knowledge search tool, have reportedly been used in disciplinary proceedings. The Times Guild, representing 1,500 editorial staff, also filed charges over the company's refusal to share information about its AI plans.

Anthropic held its Code w/ Claude London event on May 26, 2026, announcing two new Claude Managed Agents capabilities on the Claude Platform. Self-hosted sandboxes, now in public beta, let enterprises run agent tool execution on their own infrastructure or via providers like Cloudflare, Daytona, Modal, or Vercel, keeping files within their security perimeter. MCP tunnels, in research preview, allow agents to reach private Model Context Protocol servers without public internet exposure. Teams including Amplitude, Clay, and Rogo are already building on these features.

OpenAI and Thrive Holdings co-developed Tax AI using Codex for Crete's network of 30+ accounting firms, processing 7,000 tax returns this season. The system automates 1040 and 1041 preparation, saving practitioners about a third of their time and boosting throughput by 50%. A self-improving loop converts practitioner corrections into structured evals that Codex investigates and fixes autonomously. At launch, only 25% of returns hit 75% accuracy; within six weeks, 86% reached that threshold.

Runway has launched an MCP (Model Context Protocol) integration that connects its AI image and video generation tools directly into platforms like Claude, ChatGPT, Cursor, and Replit. Users can now generate polished images and videos using models including Runway's Gen-4.5, Seedance 2.0, GPT Images 2.0, and Kling without leaving their existing workflows. The move allows developers and creators to access state-of-the-art generative media models from within the tools they already use.

Microsoft's MAI Superintelligence Team has launched MAI-Image-2.5, a text-to-image model that debuted at number three on the Arena leaderboard. The model delivers significant improvements over its predecessor MAI-Image-2, including sharper text rendering, stronger stylized illustration, and better commercial and branding imagery. It also demonstrates improved visual reasoning across lighting, scale, and spatial relationships, helping turn simple directions into polished images. MAI-Image-2.5 is available now on Arena, with the MAI Playground and Azure Foundry access expected within two weeks.

Figure has signed a commercial agreement with Catalyst Brands, the operator of JCPenney, Aéropostale, and Brooks Brothers, to deploy humanoid robots at scale. The partnership will begin with an initial deployment in Reno, Nevada. Figure, founded by Brett Adcock, continues to expand its commercial robotics footprint as demand for humanoid automation in retail and logistics grows. This deal marks another step toward large-scale real-world deployment of Figure's humanoid robot platform.

Universal Music Group and TikTok have renewed their global licensing agreement, committing to remove unauthorized AI-generated music from the platform and improve artist and songwriter attribution. The deal marks a significant shift after tensions peaked in 2024, when UMG temporarily pulled its entire music catalog from TikTok over concerns about AI music and copyright violations. The agreement could serve as a template for how other platforms handle AI-generated content as regulators in the EU and U.S. tighten oversight.

ElevenLabs has launched Music v2, an updated AI music generation model featuring improved vocals, instrumentation, and arrangement across every genre. The release also brings enhanced multilingual support and new capabilities described as previously impossible with the prior version. The upgrade represents a significant step forward for AI-generated music quality and accessibility across languages, expanding the creative possibilities available to users of the ElevenLabs platform.

Datacurve has released DeepSWE, a new benchmark designed to evaluate agentic coding models more accurately than existing public leaderboards. While top models often appear closely matched in capability on standard benchmarks, DeepSWE aims to reveal where they actually diverge by reflecting the realistic day-to-day experience of developers. The benchmark sets a new standard for assessing agentic coding performance in practical, real-world conditions.

OpenAI CEO Sam Altman said at a Commonwealth Bank of Australia conference in Sydney that AI is unlikely to cause a global jobs apocalypse, admitting he was wrong about near-term economic impacts. Altman said he expected more entry-level white-collar jobs to have been eliminated by now but acknowledged a persistent human element in many roles limits AI's reach. He noted that even his own experiment using AI to reply to Slack and email messages made him reconsider full automation.

xAI has launched Grok Build in beta for all SuperGrok and X Premium+ subscribers. The new tool includes Plan Mode for structured task planning, an Imagine feature for generating images and videos, and a CLI for building automations and orchestrators. Users can access Grok Build at build.grok.com. The release expands xAI's developer and creative tooling, giving paying users on both subscription tiers access to agentic and multimedia capabilities.

Anthropic co-founder Dario Amodei joined Pope Francis at the Vatican to unveil Magnifica Humanitas, the first official Catholic teaching on artificial intelligence. The two-hour ceremony marked a significant moment in religious institutions engaging with AI ethics and governance. The document represents the Catholic Church's formal position on AI development and its implications for humanity, signaling growing interest from major global institutions in shaping moral frameworks around emerging AI technologies.

Apple registered the subdomain genai.apple.com two weeks before WWDC 2026 opens June 8, signaling new generative AI infrastructure. iOS 27 is expected to debut a rebuilt Siri powered by Google Gemini running on Apple's Private Cloud Compute, with personal context awareness, on-screen understanding, and a new Search or Ask gesture tied to the Dynamic Island. Apple also settled a $250 million class action lawsuit alleging it falsely advertised Siri capabilities on iPhone 15 Pro and iPhone 16 devices.

Anthropic co-founder Chris Olah spoke at the Vatican on May 25, 2026, at the presentation of Pope Leo XIV's AI encyclical "Magnifica humanitas: On safeguarding the human person in the time of artificial Intelligence." Olah acknowledged that frontier AI labs, including Anthropic, face commercial and geopolitical pressures that can conflict with doing the right thing, and called for outside moral voices. He highlighted three key concerns: AI's impact on the global poor, human flourishing, and the mysterious internal states researchers find inside AI models.

OpenAI has announced its first media partnership in Brazil, teaming with Grupo Folha and Grupo UOL to bring journalism from Folha de S.Paulo and UOL into ChatGPT. The deal gives more than 900 million weekly active ChatGPT users access to summaries with attribution and links to original sources. Brazil has over 50 million monthly active ChatGPT users exchanging around 140 million messages daily. Both publishers will also gain access to Codex, ChatGPT Enterprise, and the API to develop new products.