AI Reporter – June 2026

AI Update

AI-driven cybersecurity risks rapidly moved to the forefront of government and financial sector concerns. U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened major Wall Street institutions to discuss Anthropic’s Claude Mythos model and the broader threat posed by increasingly autonomous AI systems capable of identifying and exploiting software vulnerabilities at scale. The urgency intensified after Anthropic launched Project Glasswing, a restricted-access initiative involving the Cybersecurity and Infrastructure Security Agency (CISA) and major corporations, designed to uncover critical infrastructure weaknesses before public deployment. Separately, the International Monetary Fund (IMF) warned that AI-enabled cyberattacks could destabilize the global financial system because banks, utilities and telecommunications providers increasingly rely on interconnected cloud infrastructure potentially vulnerable to cascading failures.

Copyright disputes also continue to expand, with publishers and authors filing new lawsuits accusing AI developers of training models on pirated books and articles sourced from “shadow libraries.” A Northern District of California judge allowed key portions of authors’ infringement claims against Nvidia to proceed based on allegations that its Megatron model was trained using datasets containing copyrighted works.

Lawmakers, regulators and technology companies are increasingly converging around consumer protection and governance measures aimed at limiting AI-related harms before federal standards fully emerge. Pennsylvania filed what it described as a first-of-its-kind enforcement action against Character.AI, alleging the company’s chatbots unlawfully presented themselves as licensed medical professionals and misled users into believing they were receiving legitimate healthcare advice. YouTube expanded its AI likeness-detection tools to help adults identify and remove unauthorized deepfakes using their faces, reflecting growing pressure on platforms to police synthetic content.

These and other stories appear below.

AI in Business

Pentagon turns to tech companies for classified AI deployments amid Anthropic legal clash

The Department of Defense signed agreements with Nvidia, Microsoft, Amazon Web Services and Reflection AI to deploy their AI technologies on classified networks for lawful operational use. The move follows a legal dispute with Anthropic, which is currently ongoing in court. The Pentagon sought unrestricted use of Anthropic’s AI models, but Anthropic insisted on guardrails to prevent use for domestic mass surveillance and autonomous weapons. In March 2026, Anthropic secured an injunction against the Pentagon’s attempt to label it a supply-chain risk.

Source: TechCrunch

Claude Mythos and Project Glasswing: Escalating AI security risks; urgent case for governance

The release of Anthropic’s Claude Mythos Preview model— which demonstrates advanced agentic AI capabilities, such as autonomous multistep attacks and exploit generation—heightened concerns about AI security risks. In response, Anthropic initiated Project Glasswing, a coalition with restricted access for the U.S. Cybersecurity and Infrastructure Security Agency and major U.S. corporations, to identify and address critical vulnerabilities before public release. The emergence of Mythos highlights the need for robust AI governance and regulation, as current models can exhibit aggressive, profit-driven behaviors.

Source: Fortune (sub. req.)

Jensen Huang says AI is creating jobs and powering America’s reindustrialization

During a discussion at the Milken Institute, Nvidia CEO Jensen Huang addressed concerns about AI’s impact on employment, emphasizing that AI is generating new jobs rather than causing mass unemployment. Huang argued that AI represents the best opportunity for reindustrialization, as the industry relies on new types of industrial factories producing critical hardware infrastructure. Huang claims that these factories and the broader AI sector require workers, and automation of specific tasks does not equate to the elimination of entire jobs.

Source: TechCrunch

U.S. Treasury Fed leaders urge banks to use AI to strengthen cybersecurity amid rapid LLM advances

U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell recently met with Wall Street executives and major banks like JPMorgan Chase and Bank of America to discuss concerns about Anthropic’s Mythos AI model and its implications for cybersecurity. Bessent emphasized the rapid advancement of large language models and the need for American banks to proactively identify vulnerabilities in their defenses using AI tools like Mythos. He highlighted the importance of balancing safety and innovation but did not specify new regulatory measures or government actions.

Source: PYMNTS

Meta introduces AI to detect underage users on Facebook and Instagram

Meta will deploy an AI system that scans photos and videos for visual clues—such as height and bone structure—to estimate whether users are under 13 years old on Facebook and Instagram. The company emphasizes that this is not facial recognition, as the AI does not identify specific individuals but instead looks for general age-related cues. This system, already active in select countries, is part of Meta’s broader efforts to remove underage users by analyzing profiles for contextual clues (e.g., birthday celebrations, school grade mentions) across posts, comments and bios. If the AI suspects a user is underage, the account will be deactivated pending age verification. Meta plans to expand this technology to more features and countries.

Source: TechCrunch

Security researchers at RedAccess found that AI-powered software development tools, such as Lovable, Replit, Base44 and Netlify, are enabling the creation of thousands of web applications with little to no security or authentication. Over 5,000 analyzed apps were found to be accessible to anyone with the URL, and about 40% exposed sensitive data, including medical information, financial data, corporate presentations and customer chatbot logs. This highlights a significant risk as AI-driven coding tools are rapidly adopted across business sectors, potentially leading to widespread data leaks and privacy breaches.

Source: Wired (sub. req.)

IMF warns AI cyberattack threatens global financial crisis

The International Monetary Fund (IMF) issued a warning that AI-powered cyberattacks could destabilize the global financial system, especially due to the financial sector’s reliance on shared cloud services. The IMF highlighted the risks posed by Anthropic’s new AI model, Mythos, which can identify and exploit software vulnerabilities at scale, even by nonexperts.

This raises concerns not only for financial institutions but also for other sectors sharing digital infrastructure, such as energy and telecommunications. The IMF’s statement follows remarks by the Bank of England governor, who cautioned that Mythos could significantly increase cyber risk. In response, Anthropic launched Project Glasswing, providing Mythos to 40 critical companies to enhance their cyber defenses, including Nvidia, Apple, Amazon Web Services and Microsoft.

Source: Computer Weekly

YouTube is expanding its AI deepfake detection tool to all adult users

YouTube is expanding its AI likeness detection tool to all users over 18, allowing individuals to scan their faces and monitor the platform for AI-generated deepfakes using their likeness. If a match is found, users can request removal, with YouTube evaluating takedown requests under its privacy policy. Criteria for removal include whether the content is realistic and labeled as AI-generated, and if the person is uniquely identifiable, with exceptions for parody or satire. The tool only covers facial likeness and not other identifiers, like voice.

Source: The Verge

Elon Musk becomes first trillionaire after SpaceX IPO

Elon Musk is the world’s first trillionaire following a record- breaking SpaceX stock market debut that valued the company at over $2 trillion. Shares surged on opening, boosting Musk’s net worth to about $1.11 trillion, driven largely by his significant ownership in SpaceX and Tesla. The milestone has intensified debate over wealth inequality, with critics arguing such vast fortunes highlight the need for stronger taxation and regulation. Despite the headline figure, most of Musk’s wealth exists on paper, tied up in stock he cannot immediately sell. SpaceX’s IPO raised $75 billion and is expected to enrich thousands of employees, but the company remains unprofitable and relies heavily on future growth expectations.

Source: BBC

To read this Newsletter in full, please click here.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.