AI News for July 9, 2026

EU General Court ruling blocks Google's legal defense, forcing it to open Android AI layer to third-party assistants within 18 days.

The Washington Report covers recent US AI policy actions: a voluntary pre-release framework for frontier AI models, a national security memo accelerating military AI adoption, the Great American AI Act, and state-level audit requirements.

The Council of the EU has approved the Digital Omnibus on AI, updating the EU AI Act with new deadlines for labelling, high-risk systems, and prohibitions on intimate imagery and CSAM.

The US government forces OpenAI to limit GPT-5.6 Sol's rollout to 20 approved companies, marking a historic shift toward state control over powerful AI models.

Campaign staffers continue to attempt insider trades on prediction markets like Kalshi, which uses FEC data to block some bets but faces gaps in coverage.

AI companies are spending millions on the 2026 midterm elections through PACs to shape federal AI regulation, following the crypto industry's successful playbook.

A federal court ruled that AI-generated documents are not protected by attorney-client privilege, reshaping legal AI governance.

New York Times-led group asks federal court to sanction OpenAI for allegedly lying about its ability to search for copyrighted content used in AI training.

Multiple news outlets, including The New York Times and Daily News, are asking a federal judge to sanction OpenAI for alleged evidence hiding in a copyright infringement lawsuit over AI training data.

Also reported by: thebusinessjournal.com · spectrumlocalnews.com · spectrumlocalnews.com

India's Supreme Court warns against AI legal research after fabricated citations, sparking debate on regulating AI use in courts.

Iran strikes a US base in Jordan, while the US hits Iranian targets, escalating the ongoing conflict.

The article examines the opaque and ad hoc government approval process for releasing advanced AI models like OpenAI's Sol, revealing uncertainty and lack of clarity among experts and policymakers.

Illinois Governor JB Pritzker signed an AI regulation bill requiring transparency and accountability for large AI models, modeled after California and New York laws.

The New York Times and Daily News accuse OpenAI of hiding evidence and obstructing discovery in the copyright lawsuit by concealing its ability to search training data and chat logs.

The UK government unveiled a new initiative to integrate agentic AI into national cybersecurity, including the Cyber Shield project and Cyber Resilience Pledge.

Oregon's new disclosure law will be tested by the use of AI-generated political advertisements in upcoming campaigns.

North Carolina advances bipartisan HB 301 to regulate AI use in K-12 schools, mandating literacy standards and teacher training.

The article explains how the EU AI Act will require advertisers to disclose AI-generated voices in ads starting August 2, 2026, raising new production challenges.

Senators reintroduce the AI Labeling Act to require labels on AI-generated content, sparking First Amendment debate.

A California court struck down regulations that would have banned cardrooms from offering blackjack, which could have led to major job losses and economic impacts.

Weekly update on AI-related bills across states, including Illinois signing an AI Safety Act and California's legislative progress.

California and other states' laws targeting AI deepfakes in campaigns have been ineffective due to First Amendment protections and enforcement difficulties.

Bipartisan House lawmakers urge CISA, DOJ, FEC, and DHS to prepare for AI-driven threats to midterm election integrity, citing risks from chatbots and biased information sources.

China is advancing an AI for Science strategy, backed by government policies and supercomputing infrastructure, to accelerate scientific breakthroughs despite chip limitations.

Also reported by: cen.acs.org

AI voice fraud is primarily a compliance problem, requiring robust consent tracking and audit trails across channels to satisfy regulators like the FCC.

The House Foreign Affairs Committee debates the BIS budget amid corporate lobbying for AI defense funding.

An analysis of the U.S. export control order targeting Anthropic's AI models, exploring the limits of weaponized interdependence when the controlled node is a global cloud service.

A monthly roundup of global digital policy changes in June 2026, covering content moderation, AI regulation, competition policy, and data governance across G20 countries.

OpenAI launches GPT-5.6 globally with three variants, highlighting a new phase in the AI race with government oversight and Chinese competition.

OpenAI launches GPT‑Live, a new generation of voice models that enables more natural, continuous conversations by using a full-duplex architecture and delegating complex tasks to frontier models.

A detailed timeline and overview of OpenAI's GPT models and ChatGPT releases, culminating in the June 2026 preview of GPT-5.6 Sol, Terra, and Luna.

Meta releases a new coding-focused AI model, Muse Spark 1.1, with aggressive pricing to compete with OpenAI and Anthropic.

AI unicorn Mercor acquires simulation startup Deeptune, which founder Brendan Foody angel-invested in months prior, as part of its strategy to build reinforcement learning environments for frontier labs like Anthropic and OpenAI.

xAI launches Voice Agent Builder, a no-code platform for creating production voice agents using Grok Voice, with features like telephony, knowledge retrieval, tools, and guardrails.

Nvidia's data-center business continues to surge in 2026, driven by demand for AI infrastructure, as the company positions itself as the default systems vendor for AI factories.

SpaceXAI releases Grok 4.5, a cheaper alternative to Claude Opus, with integrated Cursor for coding, ahead of OpenAI's GPT-5.6 launch.

A roundup of recent AI model announcements from Anthropic, OpenAI, and Google, including Claude Sonnet 5, GPT-5.6 Sol, and Gemini Nano Banana 2 Light.

NVIDIA releases a compressed hybrid MoE LLM, Nemotron-Labs-3-Puzzle-75B-A9B, that achieves 2.03x server throughput at matched user throughput through iterative puzzle-based architecture search.

Alibaba's Qwen team releases Qwen-Image-2.0, a 7B parameter image generation model with native 2K resolution and unified generation-editing.

Alibaba Cloud forecasts 45% AI-driven revenue growth, but the China National Intelligence Law poses an unpriced legal risk for investors.

Perplexity AI is building a new coding assistant called Teammate to rival OpenAI and Anthropic, marking its expansion into developer tools.

Aissist.io's 2026 benchmark reveals AI customer service resolution rates vary widely by industry, with agentic AI outperforming retrieval-based bots, and warns that honest measurement is critical for buyers.

AI readiness is becoming a key factor for parents when choosing universities, as institutions must prepare students for an AI-driven economy.

A roundup of major tech stories including Meta's in-house AI chip, OpenAI's GPT-5.6 public rollout, Apple's EU legal loss, and startup funding news.

Anthropic releases Claude updates including monthly recap, focus settings, Microsoft 365 write tools, Cowork expansion, and government desktop beta.

A guide listing 12 verified ways to access ChatGPT Plus for free or at a lower price, including official trials, retention offers, and sharing services.

A vendor blog post from Lyzr.ai that explains what agentic AI means in banking and outlines current deployment challenges and ROI use cases.

Monica Burns discusses the digital fluency gap and shares CodeAI's free resources for teaching students to understand, question, and create with AI.

OpenAI introduces ChatGPT Work, a new desktop app, and Sites beta, while retiring group chats and Atlas in July 2026.

A study examines reproducibility in quantum computing papers, finding low code availability and executability issues.

This article compares Perplexity AI and NotebookLM as research tools, arguing they serve complementary roles in content workflows.

This guide explains how agentic AI in finance autonomously executes workflows and decisions, driving efficiency and competitive advantage for financial institutions.

A comprehensive guide comparing all Claude models from Claude 3 to Fable 5, including specs, pricing, benchmarks, and guidance for selecting the right model.

Claude's release notes detail new features like monthly recap, Claude Cowork expansion, Microsoft 365 write tools, model restorations, Sonnet 5 launch, and enterprise admin controls.

The recruitment industry is focusing on hard-to-fill AI and human-centric roles to survive automation, as demand for specialized positions surges while routine hiring becomes commoditized.

Kaiser nurses report that AI-driven surveillance and performance metrics penalize compassionate care, threatening patient safety and leading to contract negotiations.

Also reported by: calmatters.org

A Forbes article by William Arruda discusses a study linking remote work to increased social isolation and poorer mental health, offering strategies to mitigate these effects.

AI is raising client expectations for human accountants, demanding faster service, better advice, and stronger data privacy, cybersecurity, and transparency.

Issue 22 of The Skills Agenda newsletter reports on a $162M US pay-for-performance apprenticeship program and global workforce policy updates.

The article argues that the next productivity race will be won by leaders who manage AI agents as deliberately as they manage human workers.

The article argues that AI hasn't broken hiring but has exposed the existing weakness of interview-based signals, advocating for skills-based assessments.

The article critiques inadequate government and tech industry responses to potential AI-driven job losses, highlighting political proposals like Bernie Sanders' AI wealth fund and Silicon Valley's self-serving interests.

A Brown University professor's data showing a dramatic drop in student grades after switching from take-home to in-person exams highlights rising AI cheating concerns in academia.

A skeptical analysis arguing that AI-driven job losses are likely overblown based on historical precedent of technological adaptation.

Over $130 billion in AI data center projects were blocked or delayed in early 2026, highlighting a shift in constraint from hardware and power to community approval, with Bitzero positioning itself as a solution by securing permits and cheap renewable energy in Nordic regions.

GitHub launched Copilot desktop app to all users including free tier, added Codex agent support to JetBrains IDEs, and introduced enterprise usage controls.

A detailed comparison of GitHub Copilot, Cursor, and Windsurf for AI-assisted coding, evaluating features, pricing, and performance to help developers choose the best tool.

GlobalData's AI research analyst Ava is being integrated into Microsoft 365 Copilot using the Model Context Protocol, allowing business users to query proprietary market data directly within Office apps.

IBM announced new multi-agent capabilities, cost analytics, and legacy modernization packages for its IBM Bob AI development platform.

The article discusses how embedded finance, powered by AI agents, is becoming a strategic competitive advantage for enterprises across various industries.

The article analyzes recent acquisitions by Temenos, FIS, and Backbase, arguing that AI orchestration is becoming the key strategic layer in banking technology, moving beyond automation to end-to-end journey orchestration.

A live comparison table tracking pay-as-you-go API token prices for 128 AI models from 12 providers, updated daily.

A roundup of recent AI-powered marketing technology product launches and acquisitions, including Zoom's purchase of Common Room.

The article surveys multiple fronts of AI backlash in the US, including Hollywood unions, data center neighbors in Virginia, copyright lawsuits against OpenAI, and algorithmic management in warehouses and gig work.

Brown University faculty and students express worry that generative AI is harming cognitive skills and enabling cheating, following an exam scandal.

The growing societal distrust of AI, exemplified by resistance to data centers and calls for regulation, poses a significant threat to the technology's development and potential benefits.

The article discusses global debates and national policies on restricting generative AI use in schools to protect children's learning and safety.

A student argues that AI detectors are unreliable, leading to false accusations and harmful consequences in schools.

Private equity firms Vista Equity Partners and Quinti Capital made a 50% premium takeover bid for Criteo, testing private equity's continued interest in ad tech amid regulatory developments and market shifts.

New York City venture capital hit a record $4.70 billion in June 2026, driven by frontier AI labs and strong late-stage funding.

Palantir stock fell 1.6% after a report raised concerns that Democrats could target its government contracts, ending a seven-day rally.

PitchBook's Q2 2026 Venture Monitor shows record VC activity driven by megadeals and AI, with extreme concentration at the top and contraction elsewhere.

Global AI in pharma market projected to reach $28.63b by 2034 driven by drug discovery applications.

This paper reports the empirical observation of grokking and epoch-wise double descent in a two-qubit quantum neural network, identifying late-stage generalization decay and proposing weight-norm regularization as a mitigation.

A new compiler uses Möbius inversion to map diagonal quantum gates onto native multiqubit controlled-phase operations in neutral-atom hardware, improving efficiency for high-order phase terms.

Pugpig's 2026 Media App Report reveals rising app engagement through rich features, but highlights a large anonymous audience and a conversion gap limiting monetization.

A newsletter listing new AI tools such as Grok 4.5, GPT-Live, and Seedream 5.0 Pro.