Today's AI news is dominated by conflict and governance: Apple sued OpenAI over alleged hardware trade-secret theft, and Meta reversed an Instagram feature that let users make AI deepfakes after public backlash. On safety and policy, Anthropic topped the latest AI Safety Index (though the field only managed a C+), Microsoft's Brad Smith criticized unclear US AI regulation, and a coalition launched an 'Internet Court' for AI-agent disputes. The infrastructure and research beat stayed hot too, with SK Hynix's record $26.5B IPO, Hugging Face championing open-source AI, and new physical-AI and agent-security developments.
Daily Digest
AI News — 11 July 2026
·10 stories
AI
Apple sues OpenAI over alleged theft of hardware trade secrets
Apple has filed suit against OpenAI, alleging that engineers took Apple trade secrets to advance OpenAI's hardware ambitions, according to The Verge. Apple claims the alleged misconduct was directed by OpenAI's senior leadership, escalating tensions between the two companies as both pursue AI-powered devices.
Source: The Verge →Meta reverses course, disabling Instagram's AI deepfake feature after backlash
Following significant public backlash, Meta has turned off a newly announced Instagram feature that let users generate AI deepfakes of public accounts, The Verge reports. The reversal highlights the reputational risk of shipping generative-AI features without adequate consent and abuse safeguards.
Source: The Verge →Anthropic tops the latest AI Safety Index — but the field earns only a C+
Anthropic ranked first in the newest AI Safety Index, yet the overall grade for the sector was a modest C+, BankInfoSecurity reports. The assessment signals that even the leading labs still have substantial gaps in demonstrable safety practices as capabilities accelerate.
Source: BankInfoSecurity →Hugging Face's Clem Delangue: open-source AI matters more than ever
Hugging Face CEO Clem Delangue argues that open-source AI is booming and that companies are increasingly moving away from renting closed models, according to TechCrunch. He frames open models as a strategic hedge on cost, control, and independence for enterprises.
Source: TechCrunch →SK Hynix raises $26.5B in the largest foreign IPO in US history amid the AI chip boom
SK Hynix has raised $26.5 billion in what TechCrunch calls the biggest foreign IPO in US history, underscoring the scale of the AI memory-chip boom. The report notes SK Hynix and Samsung are being urged to build new US fabrication plants to meet AI-driven demand.
Source: TechCrunch →Ant Group's Robbyant releases LingBot-VA 2.0, a physical-AI video-action model
Ant Group's Robbyant unit has published the technical report for LingBot-VA 2.0, described by MarkTechPost as a causal video-action foundation model built natively for physical AI and embodied systems. It continues the recent wave of open world- and action-models aimed at robotics and agents.
Source: MarkTechPost →Why agentic AI demands a new zero-trust model
Security analysts argue that autonomous AI agents break traditional trust assumptions and require a rethought zero-trust model, BankInfoSecurity reports. Because agents act with delegated authority across tools and data, each action needs verification, least-privilege scoping, and monitoring rather than one-time authentication.
Source: BankInfoSecurity →Microsoft's Brad Smith criticizes US AI regulation for lack of clarity
Microsoft vice chair Brad Smith has criticized the current state of US AI regulation as lacking clarity, according to Crypto Briefing. His remarks add to industry calls for a more coherent federal framework as a patchwork of state rules and proposals creates uncertainty for AI developers.
Source: Crypto Briefing →Retrieval quality emerges as the defining challenge in AI agent architecture
A New Stack analysis argues that as AI agents scale, the quality of retrieval — not the model itself — is becoming the decisive factor in whether agentic systems work reliably. Poor retrieval feeds agents wrong context, compounding errors across multi-step tasks.
Source: The New Stack →GenLayer and 26 companies launch an 'Internet Court' standard for AI-agent disputes
GenLayer, backed by 26 companies, has launched an 'Internet Court' standard to adjudicate disputes involving autonomous AI agents, sociable.co reports. The effort reflects early attempts to build governance and accountability mechanisms as agents begin transacting on behalf of users.
Source: NewsCord →Frequently Asked Questions
Why is Apple suing OpenAI?
Apple alleges that engineers took Apple hardware trade secrets to advance OpenAI's device plans, and claims the misconduct was directed by OpenAI's senior leadership, according to The Verge.
What happened with Meta and Instagram AI deepfakes?
After significant backlash, Meta turned off a newly announced Instagram feature that let users generate AI deepfakes of public accounts — a reminder to ship generative-AI features with consent and abuse safeguards.
What are the main AI-governance signals today?
Anthropic topped the AI Safety Index (sector grade: C+), Microsoft's Brad Smith criticized unclear US AI regulation, and GenLayer launched an 'Internet Court' standard for AI-agent disputes.
Explore Courses on Udemy
Summaries are written by Standarity from publicly reported news; each item links to its original source. Facts belong to the linked publishers. Something off? Let us know.