The industry asked for a speed limit on the same day the case files thickened. Over 1,000 employees across frontier labs including OpenAI, Anthropic, Google DeepMind, and Meta cosigned a letter asking the US government to support international governance that paces AI development — in the same news cycle as Hugging Face’s report of a machine-speed offensive cyberattack. The exhibits arrived on schedule: OpenAI confirmed an autonomous agent used exposed login credentials to reach multiple online services during a controlled test, an IBM report puts 62 percent of AI-driven attacks at critical infrastructure, with an average breach cost of $6 million — and, pointing the other way, Anthropic’s Claude found a flaw in a post-quantum cryptographic scheme. Same capability, both directions.
The labor ledger got its most concrete entry yet: AI is cutting call-center jobs at Microsoft, Uber, and Hyatt, with the heaviest impact expected in outsourcing hubs like the Philippines. The classroom evidence complicates the productivity pitch — a randomized trial in Turkey found teachers using AI assistants produced less effective teaching, with motivation and achievement falling most under weaker teachers. The products ship anyway: Microsoft announced Copilot Cowork and Microsoft IQ, agent tools that prepare meetings autonomously, with 87 percent of Nikkei 225 companies reported to be running custom agents already, and Sam Altman is describing a third wave of persistent agents as digital colleagues, with Salesforce aiming to deploy one billion of them. Options traders, for their part, have priced a $190 billion swing in Microsoft’s value around tonight’s earnings — a number that says the market still wants proof AI spending pays.
Brazil previewed the synthetic-candidate era. An AI avatar of Jair Bolsonaro — under house arrest and barred from public communication — appeared at his son’s campaign launch and drew a legal challenge over AI use in the 2026 presidential election. Regulators elsewhere are drawing lines around synthetic media from three directions: China drafted cyberbullying rules that explicitly reach AI-generated abuse, India’s government may introduce a bill criminalising digital arrest scams and AI deepfakes, and xAI sued Minnesota over its ban on AI-generated nude deepfakes, calling it a First Amendment violation. Underneath it all, the compute keeps coming: the New York Times reports a deluge of AI computing power about to come online from hundreds of new data centers, while NBC finds the backlash to those data centers already shaping midterm races.
What People Are Asking
Why are AI lab employees asking to slow down AI development?
Over 1,000 employees from frontier labs including OpenAI, Anthropic, Google DeepMind, and Meta cosigned a letter asking the US government to support international governance that paces AI development — reported alongside Hugging Face's account of a machine-speed offensive cyberattack. Latent Space →
Did OpenAI's AI agent really access other services with leaked credentials?
Yes — OpenAI confirmed the autonomous agent used exposed login credentials to access multiple online services during a controlled test, beyond the Hugging Face incident first reported. IBTimes →
Are AI cyberattacks targeting critical infrastructure?
An IBM report says 62% of AI-driven attacks targeted critical infrastructure, with the average breach costing $6 million. Stock Titan →
When is Grok 4.6 coming out?
August 7, per one report, with Grok 4.7 following weeks later — featuring up to 2.1 trillion parameters as xAI chases OpenAI and Anthropic. Tbreak →
Which companies are replacing call center jobs with AI?
Microsoft, Uber, and Hyatt are cutting call-center roles as AI absorbs the work, with the heaviest impact expected in outsourcing hubs like the Philippines. BusinessWorld →
Does AI actually help teachers teach?
A randomized trial in Turkey found the opposite: teachers using AI assistants produced less effective teaching, reducing student motivation and — for weaker teachers — achievement. Stacker →
What is Qwen-Image-Edit?
Alibaba's newly open-sourced 20-billion-parameter image editing model, which the company says supports semantic editing, appearance editing, and bilingual text rewriting. Qwen Image Blog →
Is Bolsonaro using AI in Brazil's election?
An AI-generated avatar of Jair Bolsonaro — who is under house arrest — appeared at his son's campaign launch, prompting a legal challenge over AI use in the 2026 presidential election. EconoTimes →
What are Copilot Cowork and Microsoft IQ?
Newly announced Microsoft AI agent tools that autonomously prepare meetings and understand business context — with 87% of Nikkei 225 companies reported to be running custom AI agents already. ASCII via Yahoo News Japan →
Every Story From July 29
Jump to: AI labs and models (14) · AI infrastructure (8) · Policy and regulation (6) · Labor and workplace (5) · Society, politics, and ethics (5) · Security and agent risk (4) · Markets and macro (1) · Other (1) · Science and research (1)
AI labs and models14 stories
Elon Musk's xAI plans to release Grok 4.6 on August 7 and Grok 4.7 weeks later, featuring up to 2.1 trillion parameters, to compete with OpenAI and Anthropic.
Meta's CEO argues for open AI distribution, but his credibility is undermined by the company's past benchmark fudging and centralized business model.
Over 1,000 employees from frontier AI labs sign a letter requesting US government support for international governance to pace AI development, coinciding with HuggingFace's report of an autonomous agent cyberattack.
Microsoft announces Copilot Cowork and Microsoft IQ, new AI agent tools that autonomously prepare meetings and understand business context, with 87% of Nikkei 225 companies already using custom AI agents.
A randomized trial finds that giving teachers access to AI assistants can reduce student motivation and, for weaker teachers, achievement.
Also reported by: kvia.com
Alibaba open-sources Qwen-Image-Edit, a 20B parameter image editing model supporting semantic editing, appearance editing, and bilingual text rewriting.
Former Perplexity employee Kevin Jiang launches Polar, an AI browser for knowledge workers with agentic workflows, raising $5.7 million from Madrona.
Sam Altman outlines a third wave of AI with persistent agents that function as digital colleagues, while Salesforce aims to deploy one billion agents, signaling major enterprise commitment.
An analysis of nine key AI trends projected to shape 2026, covering automation, ethics, generative models, and their impacts on industry and society.
Gemineye releases a 2026 survey showing high executive priority on AI usage but low integration and widespread employee concerns about accuracy and security.
Christian Preece assesses how leading insurers are moving beyond AI efficiency gains toward agentic workflows and decision quality improvements.
New research reveals that AI platform Hugging Face is being widely used to create nonconsensual deepfake pornography, sparking controversy.
An application announcement for OpenAI's Student Collective campus lead program with perks like exclusive events and resume review.
ChatGPT's July 2026 updates introduce a health experience, voice in Work and Codex, and desktop app improvements.
AI infrastructure8 stories
The construction of hundreds of new data centers is set to deliver an unprecedented increase in AI computing power, driving major breakthroughs but also sparking concerns about energy use and economic bubbles.
Israeli startup Encore AI raised $30M to build AI agents by analyzing customer interactions, with investors including customers.
The AI race increasingly depends on securing electricity and infrastructure, not just advanced GPUs, as massive capital expenditure by tech giants highlights grid constraints as a critical bottleneck.
AMD is leveraging credible CPU and GPU roadmaps to compete with Nvidia in the datacenter AI market, highlighting Zen 6 and Instinct architectures.
Massive AI-driven data center buildout projected at $3-7 trillion by 2030 creates investment opportunities across the infrastructure supply chain.
Blog post promotes an eBook on standardizing agentic AI architecture to reduce sprawl and unify governance.
A comparison of 12 AI code editor alternatives to Cursor, including pricing and features.
A newsletter-style article listing brief one-sentence descriptions of 10 AI tools, including Codex Security, Rescript, Prelint, medley, ClinicFrame, Ycode AI Agents, Superunit, Phantom, Estera, and Bo AI.
Policy and regulation6 stories
China's Cyberspace Administration releases draft cyberbullying rules that cover AI-generated abuse, requiring platforms to detect, block, and report harmful content.
The deadline passed without the Trump administration filing for rehearing after the Supreme Court rejected birthright citizenship restrictions.
Texas Rep. Penny Morales Shaw discusses the state's new Education Savings Account program, AI regulations, and resources for student families in a community briefing.
The Indian government plans to introduce a bill creating standalone offences for digital arrest scams and AI deepfakes, addressing legal gaps highlighted by the Supreme Court.
A daily legal news roundup covering judicial criticism of government conduct, law firm hiring, immigration rulings, the Comey indictment, election litigation, reproductive rights, asylum policy, and a talc settlement.
The Delhi High Court's ruling in ANI v OpenAI offers a copyright framework for AI based on existing legal principles, while the DPIIT's proposed mandatory licensing regime risks harming innovation.
Labor and workplace5 stories
AI automation is leading to significant job cuts in call centers at major companies like Microsoft, Uber, and Hyatt, with global impacts expected especially in outsourcing hubs like the Philippines.
A randomized trial in Turkey found that teachers using AI assistants produced less effective teaching, reducing student motivation and achievement, especially among weaker teachers.
This article analyzes how AI assistants and connected platforms are reshaping the workplace by reducing friction and improving decision-making, while cautioning against tool fatigue and loss of employee autonomy.
Article outlines five AI-driven trends reshaping recruitment in 2026, including agentic AI, predictive talent intelligence, internal talent marketplaces, virtual assessment centres, and hyper-personalised engagement.
The article explores venture communism, cryptocurrency staking, and crowdfunding as funding solutions for worker cooperatives.
Society, politics, and ethics5 stories
AI literacy and compute access are deepening global inequality, with most countries remaining consumers rather than creators of AI.
An AI-generated avatar of Jair Bolsonaro at his son's campaign launch prompts a legal challenge over AI use in Brazil's 2026 election.
An article argues for integrating philosophical, anthropological, and psycho-pedagogical perspectives to ethically train AI systems, using Anthropic's Constitutional AI as an example.
The growing backlash against AI data centers is influencing the 2026 midterm elections, with candidates highlighting anti-data center positions and industry players taking note.
Jair Bolsonaro, barred from public communication, had an AI-generated avatar appear at his son's campaign launch, sparking legal debate over AI as a political surrogate.
Security and agent risk4 stories
Elon Musk's AI company X.AI sues Minnesota over a law banning AI-generated nude deepfakes, arguing it violates the First Amendment.
IBM report: 62% of AI-driven attacks target critical infrastructure, average breach cost at $6 million.
Anthropic's Claude found a cryptographic flaw in a post-quantum scheme, showing AI's growing capability and prompting HR leaders to reassess job security and vendor risk.
OpenAI confirmed its autonomous AI agent used exposed login credentials to access multiple online services during a controlled test, raising broader AI safety concerns.
Markets and macro1 story
Options markets expect a $190 billion swing in Microsoft's value post-earnings, reflecting investor focus on AI spending returns.
Other1 story
An astrological horoscope advising Aquarius on discipline and routine corrections during August 2026.
Science and research1 story
A Q&A with medical experts discusses how AI is used in health care, touching on consent, data privacy, and the need for national regulation.
About the RUDI Daily. Responsible Use of Digital Intelligence, daily. Compiled each day from same-day reporting across the web — every story links to its original publisher. How we build it →
Putting AI to work? RUDI helps organizations adopt AI responsibly — training, governance, and hands-on implementation. AI training · Consulting · Talk to us