AI
AI News: NY Crackdown, Apple's Hidden Siri, GPT-6 Launch & More
New York requires crawler IDs, Apple builds third-party Siri AI but hides it, OpenAI launches GPT-6 amid rivals, and NVIDIA homes data centers.
New York Passes Landmark Bill Requiring AI Crawlers to Identify Themselves
New York’s legislature has passed the New York Stealth Crawler Prohibition Act (A11292), sending it to Governor Kathy Hochul for signature. The bill requires any web crawler accessing covered news publishers to disclose its identity via a valid user-agent string and state the specific nature and purpose of its activities. Failure to comply makes the crawler a ‘stealth crawler’ and carries civil penalties up to $15,000 per day per violation. The law covers news sources that perform a public-information function, produce original content, update at least monthly, and have 1,000 or more monthly active users in New York. Enforcement rests with the state Attorney General, but publishers can also request pre-litigation subpoenas to identify violators. The bill follows research showing 80% of AI agents do not properly declare themselves on websites, and marks one of the strongest state-level efforts to regulate AI data collection.
New York passes bill forcing AI crawlers to identify themselves to news sites →
Study Finds AI-Generated Sustainability Reports More Believable Than Real Ones
Researchers at the University of Auckland Business School asked postgraduate students to use ChatGPT to generate a CEO sustainability statement for a fictional forestry company. When these AI-written reports were mixed with real corporate sustainability reports and assessed blindly by other students, the fake reports were consistently rated more credible. The students were well-versed in sustainability reporting and greenwashing, yet still favored the overly optimistic language produced by ChatGPT. The researchers found that the AI reports scored high on greenwashing metrics, glossing over real environmental trade-offs that actual companies must disclose. Human editing of the AI output did not reduce the greenwashing effect. The study raises concerns about AI’s ability to distort investor decision-making, especially as sustainability-linked investments are projected to exceed $40 trillion by 2030. The authors call for stronger regulation and auditing of sustainability disclosures to prevent AI-fabricated narratives from undermining genuine corporate accountability.
Greenwashed AI environmental reports judged more credible than the real thing →
Nearly 75% of Unemployed Workers Skip Benefits as AI Layoffs Mount
With nearly 120,000 tech workers laid off this year amid AI-driven restructurings, a key safety net is being underutilized: about 75% of unemployed individuals did not apply for unemployment insurance in 2022, a figure experts say remains accurate today. The Bureau of Labor Statistics survey found 55% believed they were ineligible, 17% expected to quickly find a new job, and others faced application hurdles. Only about 55% of applicants actually receive benefits due to state-by-state eligibility rules and employer challenges. The unemployment insurance system, designed during the New Deal, has not been updated for widespread long-term job displacement. Benefits in many states now replace only 30% of previous wages, down from the original 50% target. Columbia professor Alexander Hertel-Fernandez warns the system is ill-prepared for AI-driven layoffs, as workers may need extended time to retrain across industries. Union membership, which helps workers navigate claims, has fallen to a historic low of 9.9%. The lack of uptake among tech workers, who are more educated and likely eligible, exacerbates economic uncertainty.
Apple Built Third-Party AI Framework for Siri but Skipped WWDC Announcement
Apple’s iOS 27 developer beta includes an ‘Extensions’ framework that would allow iPhone users to switch between ChatGPT, Claude, and Gemini directly within Siri, according to Bloomberg. The system includes a settings panel and a dedicated App Store section, both toggled off on Apple’s backend. Apple has held discussions with OpenAI, Anthropic, and Google about granting entitlements, but the feature was conspicuously absent from the WWDC keynote on June 8. Three factors likely explain the omission. First, regulatory: Siri AI will not launch in the EU due to unresolved Digital Markets Act negotiations, making a third-party AI announcement awkward. Second, legal: OpenAI is preparing possible breach-of-contract action against Apple over the exclusive ChatGPT deal from June 2024, which Apple has allegedly buried behind friction. Third, messaging: Apple spent two years rebuilding Siri and wanted to promote its own AI, not a model picker. The framework is built and could launch later, but Apple’s strategy remains in flux amid these pressures.
Why Apple built a third-party AI system for Siri and then refused to show it at WWDC →
Pokémon Go Player Data Used to Train AI Now Linked to Military Drones
Millions of Pokémon Go players voluntarily scanned real-world locations with their smartphones after a 2021 in-game update incentivized the activity. Niantic used those billions of visual data points to train its spatial AI models, later spun off into Niantic Spatial. In December 2025, Niantic Spatial partnered with Vantor, a US defense intelligence firm, to combine its Visual Positioning System with Vantor’s Raptor software for GPS-free drone navigation. The system achieved error reduction up to 70% and accuracy around 1.5 meters in early tests. The AR scans themselves were not directly provided to Vantor; they were used to train Niantic’s foundation models, the company stressed. However, the partnership raises privacy and ethical questions about user-generated data being repurposed for military applications. Vantor also secured a US Army contract worth up to $217 million for 3D terrain data under the One World Terrain program. GPS jamming and spoofing are already active threats in conflicts like Ukraine and Iran, making visual navigation systems increasingly valuable for defense.
Pokémon Go data helped train AI now linked to military drones →
Microsoft CEO Nadella Warns of AI Concentration Risk and Calls for Proprietary Learning
In a blog post, Microsoft CEO Satya Nadella argued that companies must build proprietary learning systems to avoid ceding all value to a few AI models. He introduced the concept of ‘token capital’ alongside human capital, urging firms to create internal training setups, private evals, and queryable institutional knowledge. ‘You can offload a task, or even a job, but you can never offload your learning,’ Nadella wrote, emphasizing that the future firm compounds learning across people and AI. Nadella warned of a scenario where ‘a small number of AI systems capture all the economic returns, while entire industries find their knowledge commoditized right out from underneath them.’ This marks a shift from his earlier view in March 2025 that models were becoming commodities. He now fears concentration, as OpenAI and Anthropic build product ecosystems around their models. Nadella’s advice aligns with Microsoft’s strategy to keep enterprises on Azure and Office bundles, but also reflects a genuine concern about AI market dominance.
NVIDIA to Install Mini AI Data Centers in Homes via XFRA Nodes
NVIDIA is rolling out XFRA nodes, compact computing units with 16 GPUs, AMD processors, and 3TB of memory, designed for residential and small business use. The liquid-cooled, fanless systems integrate with smart electrical panels and backup batteries, using underutilized residential electricity. A pilot program with homebuilder Pulte Group and energy management firm Span will install the nodes in 100 homes across the southwestern US, with plans to scale to tens of thousands by 2027. Homeowners receive the hardware at no upfront cost and benefit from reduced utility bills, backup power, and solar optimization. NVIDIA positions the initiative as a distributed cloud network that eases grid strain and public opposition to massive data centers. However, challenges include scalability, security risks from outdoor hardware placement, and logistical complexity. If successful, the model could democratize AI computing power and reshape infrastructure deployment.
Why NVIDIA Wants to Install a Mini Data Center in Your Home →
OpenAI Launches GPT-6 Amid Rising Competition and Financial Pressure
OpenAI has released GPT-6, aiming to reclaim its competitive edge after falling behind rivals. The model addresses the ‘goblin incident’ reward model flaws seen in GPT-5.5 and expands persistent memory capabilities. However, OpenAI faces headwinds: Microsoft replaced OpenAI’s tech with its Polaris model in GitHub Copilot, Google integrated Gemini into Siri alongside Apple, and Anthropic’s Claude models outperform GPT-5.5 on benchmarks. Financially, OpenAI is pursuing a $1 trillion IPO despite a $14 billion operating loss in 2023 and no profitability expected until 2030. The company also navigates multiple lawsuits and EU AI Act compliance. The massive Stargate project aims to secure computational power equivalent to 1% of the US electrical grid to support GPT-6’s scalability. GPT-6’s success is critical to justify OpenAI’s valuation and investor confidence in a rapidly commoditizing market.
What ChatGPT 6 Means for OpenAI Now That Microsoft and Google Walk Away →