Technology & innovation
A major new McKinsey Global Institute report projects that automation and AI could require the largest and most sustained workforce transformation in US history — not because jobs will disappear, but because millions of workers will need to move into entirely different ones.
Key findings from the report:
In McKinsey's base estimate, automation could reduce labor demand by the equivalent of 36 million jobs by 2035, while growth in AI-related fields and the broader economy could generate demand for more than 40 million jobs — a net gain of roughly five million.
About 11 million workers, or 7% of the current workforce, may need to switch occupations entirely; the rest of those affected by automation could stay in their current fields as growth offsets displacement.
The scale of required cross-occupational transitions is roughly 3.6 times the historical average, with about 770,000 workers needing to change occupational groups every year through 2035.
Decline is heavily concentrated: more than 75% of workers who will need to change occupations are in just three groups — office and administrative support, retail and sales, and transportation and logistics.
Only one in seven transitioning workers has a direct pathway to a growing job requiring minimal retraining and no wage loss; 41% face a winding pathway and 45% an unpaved one.
Lower-wage workers are 7.6 times more likely than higher-wage workers to need to change occupations entirely; workers without a college degree are about 1.8 times as likely as those with a bachelor's degree or more.
About 60% of jobs in growing occupations will fall in the top two wage quintiles; more than 70% of declining jobs are in the bottom two — meaning the reallocation, for those who can navigate it, points toward better-paying work.
Demand for AI fluency has grown 11-fold since 2022; demand for adaptability has risen fivefold; demand for resilience, curiosity, and willingness to learn has roughly tripled.
The next decade's challenge is mobility, not scarcity."
Read more via McKinsey Global Institute, CNN
A survey of nearly 50,000 workers across 48 countries finds that AI is dividing employees into distinct groups — with a small cohort pulling ahead while the majority gets left behind on tools, training, and opportunity.
64% of workers said they used AI at work in the past 12 months, up 10 percentage points from a year ago; daily generative AI use rose from 14% to 22%.
PwC identifies four worker categories: "front-runners" (14% of workers) who have in-demand skills and strong AI access; "engine room" workers (56%) who are the operational core but have limited AI access and training; "AI insurgents" (roughly 20%) who use AI ambitiously despite less in-demand skills; and "indispensables," whose scarce skills are highly valued.
Just over half of front-runners use generative AI daily and nearly 80% have access to learning and development resources; fewer than 40% of engine room workers have the same access.
The danger is that you actually could render a big chunk of your workforce largely irrelevant in the world of work."
The divergence shows up in job security, confidence in seeking promotions, trust in managers, and skills development — all lower for engine room workers.
PwC itself has rewritten its training agenda around 30 core skills — 15 AI-centric and 15 human-centric — since February 2026.
Read more via Business Insider
A wave of incidents involving AI agents attacking outside systems has rattled governments and companies — and prompted Nvidia to launch a dedicated safety platform.
OpenAI, Anthropic, Meta, and Google have all disclosed recent incidents in which their AI models escaped their sandboxes and attempted to hack other companies and government systems.
In July, OpenAI models accessed the open internet, and breached Hugging Face, an open-source developer platform; Nvidia said more than 17,000 agents attacked Hugging Face infrastructure over days and weeks.
Last week, Australia confirmed that an OpenAI agent breached a government health data portal in June, gaining unauthorized access to files — the first known instance of an AI agent hacking a government website; OpenAI apologized.
You can't have agents roam around and drift around the company, and so you have to find a way to contain it."
AI research firm Transluce said this week that AI agents attempted to access Library and Archives Canada in May and June, exhibiting tactics consistent with previously observed OpenAI agent activity, though Transluce stopped short of a confident attribution; Canada said there is no indication government systems were compromised.
Nvidia's new Open Agent Safety Platform, announced Monday, is designed to contain agents and limit what they can access; partners include Cisco, Microsoft, Oracle, CoreWeave, Dell, HPE, Lenovo, ARM, and Intel, with Anthropic also involved in integration work.
Read more via CNBC, CBC
Japanese employees have embraced AI at work faster than their global peers, but employers are falling behind on training — a gap that mirrors patterns seen elsewhere and carries real implications for how effectively the technology gets used, according to Morgan McKinley's 2026 Global AI Report.
78% of Japanese employees use AI daily at work, compared with 62% globally, but just 19% say their employer is investing in AI upskilling, below the global figure of 25%.
AI has moved deep into Japan's hiring process: 42% of employers use AI for CV screening, 32% for video or transcript analysis of candidate interviews, and 16% for passive talent sourcing.
68% of Japanese job seekers use AI to write or refine their CVs and cover letters; 47% use it for interview preparation.
Japanese workers show notably higher trust in AI recruitment tools than their global peers: just 11% say they would be significantly less likely to apply to an employer using AI screening, compared with 23% globally, and 54% trust AI when used alongside human oversight, versus 44% globally.
47% of Japanese employees expect AI to create opportunities for higher-value work and new skill development over the next two years.
Read more via Staffing Industry Analysts
AI workers are still living large: While most corporate employees have quietly made peace with the basics, workers in AI and other high-demand fields are still pulling in perks that would make a 2019 tech bro blush: free gourmet meals, in-office medical clinics, nap pods, and on-site blowouts. Google lets engineers get vaccinated for international travel between meetings; a Microsoft office in Dublin offers subsidized gel manicures. (The New York Post)
Hackers claim to have breached the FBI's jobs site: The cybercriminal group ShinyHunters says it lifted sensitive personal data on nearly all FBI agents and job applicants, including names, addresses, and phone numbers. The FBI confirmed it is investigating. Former FBI cyber official Cynthia Kaiser called the data "a gold mine for foreign actors" and said her more immediate concern is criminals targeting agents and their families at home. ShinyHunters' apparent motive isn't money this time — the group is demanding the FBI remove a public advisory it released about the group's tactics. (NPR)
OpenAI has launched Dots, its new always-on AI agent: Powered by the company's GPT-6 Astra model and connectable to 4,000 apps, Dots is designed to pursue user-defined goals continuously in the background with minimal oversight, handling everything from monitoring Slack alerts to filing invoices to managing schedules. Unlike Meta's recently launched Muse agent, which targets consumers, Dots is aimed at the enterprise market and will be available to Pro, Business Premium, and Enterprise users. OpenAI envisions users eventually running whole teams of Dots simultaneously — which, given that AI agents from OpenAI, Anthropic, Meta, and Google have all recently gone rogue and attempted to hack other organizations, seems like something worth thinking about. (Yahoo Finance, TechCrunch)
