Latest AI News

Anthropic has opened a Sydney office and appointed Theo Hourmouzis, formerly Senior Vice President at Snowflake for Australia, New Zealand, and ASEAN, as General Manager for Australia and New Zealand. Hourmouzis brings over 20 years of Asia Pacific tech leadership experience and will lead Anthropic's local team, deepening enterprise relationships with Commonwealth Bank and Quantium. New partnerships include Canva, Xero, and YMCA South Australia, while Anthropic also holds an MOU with the Australian government.

OpenAI has open-sourced Symphony, an agent orchestration spec built by Alex Kotliarskyi, Victor Zhu, and Zach Brock that turns Linear issue trackers into a control plane for Codex coding agents. Every open ticket gets a dedicated agent that runs continuously until complete, with humans reviewing results. Teams using Symphony saw landed pull requests increase 500% in three weeks. The system is technically a SPEC.md file on GitHub, allowing non-engineers like product managers to file feature requests directly and receive working implementations.

Over 600 Google employees, including more than 20 principals, directors, and vice presidents, signed a letter to CEO Sundar Pichai demanding that Google block the Pentagon from using its Gemini AI models in classified settings. Organizers say many signers work at Google DeepMind. The letter warns that classified use could cause harms without employees' knowledge or ability to stop them. The move follows a report that Google and the Pentagon are in active talks over a classified Gemini deployment deal.

According to analyst Ming-Chi Kuo, OpenAI is developing a smartphone to compete with the iPhone, targeting mass production in 2028. OpenAI is working with MediaTek and Qualcomm on processors, with Luxshare as the exclusive manufacturing partner. Specifications are expected to be finalized by late 2026 or early 2027. Kuo believes AI agents will make the device feel distinct from an iPhone. OpenAI is also collaborating with former Apple design chief Jony Ive on other hardware, including a ChatGPT-powered smart speaker.

Elon Musk and Sam Altman are heading to a jury trial beginning April 27th in Northern California federal court over Musk's 2024 lawsuit accusing OpenAI of abandoning its nonprofit mission to benefit humanity in favor of profits. Musk, a cofounder, claims Altman and co-founder Greg Brockman deceived him into funding the company. He is seeking removal of both executives, up to $150 billion in damages, and an end to OpenAI's public benefit corporation status. OpenAI calls the suit a baseless attempt to harm a competitor.

Microsoft is expanding its Dynamics 365 platform with new agentic customer experience capabilities, including real-time voice agents now generally available in Copilot Studio. The update adds agents across Dynamics 365 Contact Center, Sales, and Customer Insights. Key additions include a Customer Assist Agent and Quality Assurance Agent, both generally available, plus a Service Operations Agent in public preview. In Dynamics 365 Sales, five new features include a Sales Opportunity Agent and Voice to CRM notes for Outlook. Conversational Journeys is also expanding to SMS.

Microsoft and OpenAI have amended their partnership agreement, giving OpenAI the flexibility to serve products across any cloud provider while keeping Azure as its primary and first-launch platform. Microsoft's license to OpenAI IP becomes non-exclusive through 2032, and Microsoft will no longer pay a revenue share to OpenAI. OpenAI's revenue share payments to Microsoft continue through 2030 with a total cap. Microsoft remains a major shareholder as both companies plan to scale datacenter capacity and collaborate on next-generation silicon.

OpenAI CEO Sam Altman published a principles document outlining five core commitments guiding the company's AGI development: democratization, empowerment, universal prosperity, resilience, and adaptability. OpenAI pledges to resist power consolidation, ensure democratic oversight of AI decisions, and expand global AI infrastructure to lower costs. The company also commits to collaborating with governments and international agencies on safety risks, including pathogen creation and cybersecurity threats, while iteratively deploying systems and updating its positions as the technology evolves.

Alibaba's Qwen team has launched Qwen-Image-2.Pro, a new text-to-image model that improves image quality, multilingual text rendering, and instruction following compared to its predecessor. The model also delivers more consistent performance across different visual styles. Qwen-Image-2.Pro has earned a ranking of ninth worldwide in the Text-to-Image category on Arena, a competitive AI benchmarking platform. The model is currently available to try on ModelScope.

Grok Imagine has rolled out a significant update to its image-to-video generation feature, delivering dramatically improved lip sync and sharper audio quality across all generations on the platform. The enhancement ensures that dialogue accurately tracks mouth movements while sound is matched to match scene context, resulting in more realistic and coherent video output. The update is designed to close the gap between creator intent and final output, producing videos that look and sound closer to what users originally envisioned.

Google parent Alphabet will invest up to $40 billion in AI startup Anthropic, committing $10 billion immediately at a $350 billion valuation, with $30 billion more contingent on performance targets. The deal deepens Google's partnership with Anthropic, maker of the Claude AI models, just days after Amazon announced a separate $25 billion investment. Anthropic's annual run-rate revenue has surpassed $30 billion, up from $9 billion at end of 2025, driven largely by strong developer adoption of its Claude Code tool.

Meta has partnered with Amazon Web Services to add tens of millions of AWS Graviton cores to its compute portfolio, making Meta one of the largest Graviton customers in the world. The deal addresses the growing CPU demands of agentic AI systems that reason, plan, and execute complex tasks at scale. AWS Graviton5 cores offer faster data processing and greater bandwidth for these workloads. The agreement reflects Meta's broader strategy of diversifying compute infrastructure rather than relying on a single chip architecture.

Forbes's Leadership YouTube playlist, which contains 324 videos and has accumulated nearly 35,000 views, features a range of short-form guides aimed at business leaders navigating AI adoption. Recent entries cover topics such as tasks leaders should avoid delegating to AI, how small businesses can use AI to compete with larger rivals, AI hiring considerations for 2026, and a day AI implementation playbook. The playlist also addresses AI-driven answer engine optimization as a replacement for traditional SEO strategies.

OpenAI has released GPT-5.5 via its API, offering higher intelligence and stronger token efficiency compared to previous models. The update is designed to handle complex tasks more effectively, reducing the number of retries needed to complete work. Improved token efficiency means developers can accomplish more within the same token budget, lowering costs and improving reliability for demanding applications built on OpenAI's platform.

Anthropic and NEC Corporation have announced a strategic partnership to deploy Claude to approximately 30,000 NEC Group employees worldwide, making NEC Anthropic's first Japan-based global partner. Together they will develop secure, industry-specific AI products for Japanese markets in finance, manufacturing, cybersecurity, and local government. NEC will integrate Claude, including Claude Opus 4.7 and Claude Code, into its BluStellar Scenario platform, and establish a Center of Excellence to build one of Japan's largest AI-native engineering organizations.

Anthropic has updated Claude's election safeguards ahead of the 2026 US midterms and Brazil's elections. Claude Opus 4.7 and Sonnet 4.6 scored 95% and 96% on political bias evaluations, and handled election-related policy prompts appropriately 100% and 99.8% of the time. New tests also assessed autonomous influence operations, with safeguards causing models to refuse nearly every task. Claude.ai will display election banners directing users to TurboVote, a nonpartisan resource from Democracy Works, for voting information.

Spotify has launched a Claude integration allowing users to connect their Spotify account directly inside Claude for personalized music and podcast recommendations based on their listening history. Both Free and Premium users can access the feature, with Premium subscribers able to describe a mood or vibe to generate a custom playlist. Claude also supports Spotify Connect, enabling device switching without leaving the conversation. Spotify confirmed it does not share audio content with Anthropic for training. The integration is available globally on web, iOS, Android, and desktop.

Anthropic traced Claude Code quality degradation to three issues, all resolved by April 20 (v2.1.116). On March 4, default reasoning effort was downgraded from high to medium to cut latency, hurting intelligence; reverted April 7. A March 26 caching bug caused Claude to drop prior reasoning every turn after idle sessions, making it forgetful and draining usage limits; fixed April 10. A system prompt added April 16 capping responses to 25 words caused a 3% intelligence drop and was reverted April 20. Anthropic is resetting usage limits for all subscribers.

Anthropic has expanded Claude's connector directory to include everyday lifestyle apps, adding AllTrails, Audible, Booking.com, Instacart, Intuit TurboTax, Intuit Credit Karma, Resy, Spotify, StubHub, Taskrabbit, Thumbtack, TripAdvisor, Uber, Uber Eats, and Viator. Since launching in July 2025, the directory has grown to over 200 connectors. Claude now dynamically suggests relevant apps mid-conversation based on context, and can chain multiple services together. User data from connected apps is not used to train models, and Claude confirms before making purchases or bookings.

OpenAI has launched a Bio Bug Bounty program for GPT-5.5, offering $25,000 to the first researcher who finds a universal jailbreak capable of bypassing the model's five-question bio safety challenge. Open to AI red teamers, security experts, and biosecurity researchers, the program runs from April 28 to July 27, 2026, with applications accepted through June 22. Testing is limited to GPT-5.5 in Codex Desktop only, and all participants must sign an NDA.

Leonardo.Ai Head of Creative Dwayne Koh tested GPT-Image 2.0 on advertising briefs using plain, unprompted language and found results that were creatively coherent, not just technically accurate. The model spontaneously added a TikTok sticker to a Gen Z lip balm ad without being asked. Koh, who has worked with Adidas Originals, Levi's, and Jordan, argues AI is shifting from rendering tool to taste-reflecting partner, and proposes inverting the creative process by generating ad outputs first, then reverse-engineering product identity from patterns.

Microsoft has launched hosted agents in Foundry Agent Service, now in public preview. Each agent session gets its own hypervisor-isolated sandbox with persistent filesystem, per-session Entra Agent ID, scale-to-zero billing, and predictable cold starts. Developers can bring any framework — LangGraph, OpenAI Agents SDK, Claude Agent SDK, or Microsoft Agent Framework v1.0 — and deploy with a single azd deploy command. Agents can also publish directly to Microsoft Teams and M365 Copilot.

Microsoft has made agentic capabilities in Word, Excel, and PowerPoint generally available for Microsoft 365 Copilot and Microsoft 365 Premium subscribers. Copilot can now take multi-step, app-native actions directly in documents, worksheets, and presentations rather than just suggesting steps. Early data shows significant gains: Excel engagement rose 67%, retention 50%, and satisfaction 65%. Word saw a 52% engagement increase, while PowerPoint retention climbed 36%. The features are also available to Microsoft 365 Personal and Family plan users.

Anthropic has published guidance on connecting Claude agents to production systems using the Model Context Protocol (MCP), outlining three integration approaches: direct API calls, CLIs, and MCP. MCP is recommended for production cloud-hosted agents due to its standardized auth, portability across clients like Claude, ChatGPT, and Cursor, and rich semantics. MCP SDKs recently surpassed 300 million monthly downloads. Key patterns include grouping tools by intent, using tool search to cut token usage by 85%, and pairing MCP servers with skills for complex workflows.

Anthropic has launched the Economic Index Survey, a monthly study conducted via its Anthropic Interviewer tool to track how AI is reshaping work and the economy. Each month, a randomly selected group of Claude users with accounts at least two weeks old will be invited to share how AI is changing their tasks, productivity, and hiring expectations. Results will feed into future Anthropic Economic Index reports, complementing a companion study analyzing 81,000 open-ended responses collected in December 2024.

OpenAI has launched ChatGPT for Clinicians, a free tool available to verified U.S. physicians, nurse practitioners, physician assistants, and pharmacists. Built with input from hundreds of physician advisors, it supports documentation, medical research, and care consults, and includes trusted clinical search, CME credit tracking, and optional HIPAA compliance. Physician advisors rated 99.6% of responses safe and accurate across 6,924 test conversations. OpenAI also released HealthBench Professional, an open benchmark for evaluating AI on real clinician tasks.

OpenAI has added WebSocket support to its Responses API, cutting agentic workflow latency by up to 40%. Instead of rebuilding full conversation state on each HTTP request, WebSocket mode maintains a persistent connection with in-memory caching of previous response state, tool definitions, and rendered tokens. The update also included safety stack improvements and reduced network hops, achieving a 45% faster time to first token. Codex, Vercel's AI SDK, Cline, and Cursor all reported significant speed gains, with GPT-5.Codex-Spark hitting 1,000 tokens per second.

xAI has launched Grok Voice Think Fast 1.0, its new flagship voice agent model, now available via API. Built for complex, multi-step workflows in customer support, sales, and enterprise applications, it features real-time background reasoning with zero added latency, precise data entry, and support for plus languages. The model tops the τ-voice Bench leaderboard and already powers Starlink's phone support, achieving a 70% autonomous resolution rate and 20% sales conversion rate using 28 distinct tools.

Google unveiled its eighth-generation TPUs at Google Cloud Next, introducing two purpose-built chips: TPU 8t for training and TPU 8i for inference. Designed with Google DeepMind, TPU 8t scales to 9,600 chips delivering 121 ExaFlops per superpod with nearly 3x compute performance over the prior generation. TPU 8i offers 80% better performance-per-dollar and 288 GB of high-bandwidth memory for low-latency agentic workloads. Both chips target the demands of AI agents and will reach general availability later this year.

Google has made Gemini Embedding 2 generally available via the Gemini API and Vertex AI. The natively multimodal embedding model supports search and reasoning across text, image, video, and audio data, eliminating the need for complex, fragmented pipelines. During its preview phase, developers built prototypes including advanced e-commerce discovery engines and video analysis tools. The general availability release brings production-ready stability and optimizations, and the technology also powers multiple Google products internally.

Ideogram has launched Custom Models, a new feature that lets users train image generation models on their own images rather than general internet data. The tool is designed to ensure every generated image follows a brand's specific art direction, typography, and visual identity. Custom Models are available now on Ideogram's Team, Pro, and Enterprise plans and can be accessed via API, addressing a key gap in existing image models that struggle to produce consistent on-brand visuals.

OpenAI has released Privacy Filter, an open-weight model for detecting and redacting personally identifiable information in text. The 1.5B-parameter model, with only 50M active parameters, runs locally to keep unfiltered data on-device. It achieves a 97.43% F1 score on the corrected PII-Masking-300k benchmark and supports up to 128,000 tokens of context. It detects eight PII categories including names, addresses, emails, phone numbers, and API keys. Available under Apache 2.0 on Hugging Face and GitHub.

Alibaba's Qwen team has released Qwen3.6-27B, a dense open-source model designed for coding tasks. Despite having only 27 billion parameters, the model claims to outperform the much larger Qwen3.5-397B-A17B across all major coding benchmarks, demonstrating strong agentic coding capabilities. The release highlights how a smaller dense model can exceed a significantly larger mixture-of-experts model in specialized performance, suggesting improved efficiency in Alibaba's latest generation of open-source AI.

YouTube is expanding its likeness detection tool to the entertainment industry, including talent agencies, management companies, and celebrities. Similar to Content ID, the tool scans for AI-generated deepfakes of a person's face and lets them request removal. Developed with support from major agencies including CAA, UTA, WME, and Untitled Management, the tool is now accessible to celebrities and entertainers even if they do not have a YouTube channel.

Meta is installing tracking software called Model Capability Initiative (MCI) on U.S. employees' computers to capture mouse movements, clicks, keystrokes, and occasional screenshots for AI model training. CTO Andrew Bosworth announced the effort under a rebranded "Agent Transformation Accelerator" program, aiming to build autonomous AI agents that handle work tasks. Meta spokesperson Andy Stone said data won't be used for performance reviews. Legal experts warn the practice may violate European GDPR rules and mirrors surveillance previously limited to gig workers.

X has launched Custom Timelines, one of its biggest changes to the platform's core feed experience. The new feature, powered by Grok, lets users pin specific topics directly to their home tab. With support for over 75 available topics, users can curate a focused feed around their favorite niches. The update gives users significantly more control over what content appears on their main timeline, moving away from a purely algorithmic approach to content discovery.

SpaceX has confirmed a deal giving it the option to acquire AI coding platform Cursor for $60 billion or pay a $10 billion fee for their collaboration. The New York Times first reported the arrangement, which SpaceX confirmed on X. The two companies will work together using SpaceX's Colossus supercomputer, equivalent to one million H100 GPUs. The deal comes ahead of a SpaceX IPO, with Cursor having recently sought $2 billion in funding at a $50 billion valuation.

Google is expanding benefits for AI Pro and Ultra subscribers in AI Studio, offering increased usage limits and access to new models including Nano Banana Pro and Gemini Pro. The update, announced by Product Manager Seth Odoom, lets developers move from idea to working application quickly with predictable costs. For users who've hit free-tier limits, a Google AI plan acts as a low-setup billing bridge before transitioning to pay-per-request API keys for production launches.

Google has added Continued Conversation to Gemini for Home, fulfilling one of users' top requests since early access launched. After an initial "Hey Google" command, the microphone stays active for a few seconds so users can follow up without repeating themselves. Key improvements over Google Assistant include conversational context memory, multilingual support across all supported languages and regions, smarter side-talk detection to reduce accidental responses, and whole-home access for all household members including guests.

Google has open-sourced the DESIGN.md specification from its Stitch AI UI design tool, allowing design rules to be shared across any platform or tool. DESIGN.md lets users export and import design systems between projects, so AI agents understand the reasoning behind color choices and can validate decisions against WCAG accessibility standards. Previously limited to Stitch, the open draft spec is now available on GitHub, enabling broader adoption and consistent AI-generated interfaces that match brand guidelines without rebuilding rules from scratch.